Merge branch 'ino-alloc' of git://repo.or.cz/linux-btrfs-devel into inode_numbers

Conflicts:
	fs/btrfs/free-space-cache.c

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

fs/btrfs/btrfs_inode.h

@@ -166,6 +166,15 @@ static inline struct btrfs_inode *BTRFS_I(struct inode *inode)
 	return container_of(inode, struct btrfs_inode, vfs_inode);
 }
 
+static inline u64 btrfs_ino(struct inode *inode)
+{
+	u64 ino = BTRFS_I(inode)->location.objectid;
+
+	if (ino <= BTRFS_FIRST_FREE_OBJECTID)
+		ino = inode->i_ino;
+	return ino;
+}
+
 static inline void btrfs_i_size_write(struct inode *inode, u64 size)
 {
 	i_size_write(inode, size);

fs/btrfs/compression.c

@@ -125,9 +125,10 @@ static int check_compressed_csum(struct inode *inode,
 		kunmap_atomic(kaddr, KM_USER0);
 
 		if (csum != *cb_sum) {
-			printk(KERN_INFO "btrfs csum failed ino %lu "
+			printk(KERN_INFO "btrfs csum failed ino %llu "
 			       "extent %llu csum %u "
-			       "wanted %u mirror %d\n", inode->i_ino,
+			       "wanted %u mirror %d\n",
+			       (unsigned long long)btrfs_ino(inode),
 			       (unsigned long long)disk_start,
 			       csum, *cb_sum, cb->mirror_num);
 			ret = -EIO;

fs/btrfs/ctree.h

@@ -105,6 +105,12 @@ struct btrfs_ordered_sum;
 /* For storing free space cache */
 #define BTRFS_FREE_SPACE_OBJECTID -11ULL
 
+/*
+ * The inode number assigned to the special inode for sotring
+ * free ino cache
+ */
+#define BTRFS_FREE_INO_OBJECTID -12ULL
+
 /* dummy objectid represents multiple objectids */
 #define BTRFS_MULTIPLE_OBJECTIDS -255ULL
 
@@ -830,9 +836,6 @@ struct btrfs_block_group_cache {
 	u64 bytes_super;
 	u64 flags;
 	u64 sectorsize;
-	int extents_thresh;
-	int free_extents;
-	int total_bitmaps;
 	unsigned int ro:1;
 	unsigned int dirty:1;
 	unsigned int iref:1;
@@ -847,9 +850,7 @@ struct btrfs_block_group_cache {
 	struct btrfs_space_info *space_info;
 
 	/* free space cache stuff */
-	spinlock_t tree_lock;
-	struct rb_root free_space_offset;
-	u64 free_space;
+	struct btrfs_free_space_ctl *free_space_ctl;
 
 	/* block group cache stuff */
 	struct rb_node cache_node;
@@ -1107,6 +1108,16 @@ struct btrfs_root {
 	spinlock_t accounting_lock;
 	struct btrfs_block_rsv *block_rsv;
 
+	/* free ino cache stuff */
+	struct mutex fs_commit_mutex;
+	struct btrfs_free_space_ctl *free_ino_ctl;
+	enum btrfs_caching_type cached;
+	spinlock_t cache_lock;
+	wait_queue_head_t cache_wait;
+	struct btrfs_free_space_ctl *free_ino_pinned;
+	u64 cache_progress;
+	struct inode *cache_inode;
+
 	struct mutex log_mutex;
 	wait_queue_head_t log_writer_wait;
 	wait_queue_head_t log_commit_wait[2];
@@ -2413,12 +2424,6 @@ int btrfs_del_orphan_item(struct btrfs_trans_handle *trans,
 			  struct btrfs_root *root, u64 offset);
 int btrfs_find_orphan_item(struct btrfs_root *root, u64 offset);
 
-/* inode-map.c */
-int btrfs_find_free_objectid(struct btrfs_trans_handle *trans,
-			     struct btrfs_root *fs_root,
-			     u64 dirid, u64 *objectid);
-int btrfs_find_highest_inode(struct btrfs_root *fs_root, u64 *objectid);
-
 /* inode-item.c */
 int btrfs_insert_inode_ref(struct btrfs_trans_handle *trans,
 			   struct btrfs_root *root,

fs/btrfs/disk-io.c

@@ -41,6 +41,7 @@
 #include "locking.h"
 #include "tree-log.h"
 #include "free-space-cache.h"
+#include "inode-map.h"
 
 static struct extent_io_ops btree_extent_io_ops;
 static void end_workqueue_fn(struct btrfs_work *work);
@@ -1326,6 +1327,19 @@ again:
 	if (IS_ERR(root))
 		return root;
 
+	root->free_ino_ctl = kzalloc(sizeof(*root->free_ino_ctl), GFP_NOFS);
+	if (!root->free_ino_ctl)
+		goto fail;
+	root->free_ino_pinned = kzalloc(sizeof(*root->free_ino_pinned),
+					GFP_NOFS);
+	if (!root->free_ino_pinned)
+		goto fail;
+
+	btrfs_init_free_ino_ctl(root);
+	mutex_init(&root->fs_commit_mutex);
+	spin_lock_init(&root->cache_lock);
+	init_waitqueue_head(&root->cache_wait);
+
 	set_anon_super(&root->anon_super, NULL);
 
 	if (btrfs_root_refs(&root->root_item) == 0) {
@@ -2404,12 +2418,15 @@ int btrfs_free_fs_root(struct btrfs_fs_info *fs_info, struct btrfs_root *root)
 	if (btrfs_root_refs(&root->root_item) == 0)
 		synchronize_srcu(&fs_info->subvol_srcu);
 
+	__btrfs_remove_free_space_cache(root->free_ino_pinned);
+	__btrfs_remove_free_space_cache(root->free_ino_ctl);
 	free_fs_root(root);
 	return 0;
 }
 
 static void free_fs_root(struct btrfs_root *root)
 {
+	iput(root->cache_inode);
 	WARN_ON(!RB_EMPTY_ROOT(&root->inode_tree));
 	if (root->anon_super.s_dev) {
 		down_write(&root->anon_super.s_umount);
@@ -2417,6 +2434,8 @@ static void free_fs_root(struct btrfs_root *root)
 	}
 	free_extent_buffer(root->node);
 	free_extent_buffer(root->commit_root);
+	kfree(root->free_ino_ctl);
+	kfree(root->free_ino_pinned);
 	kfree(root->name);
 	kfree(root);
 }

fs/btrfs/export.c

@@ -32,7 +32,7 @@ static int btrfs_encode_fh(struct dentry *dentry, u32 *fh, int *max_len,
 	len  = BTRFS_FID_SIZE_NON_CONNECTABLE;
 	type = FILEID_BTRFS_WITHOUT_PARENT;
 
-	fid->objectid = inode->i_ino;
+	fid->objectid = btrfs_ino(inode);
 	fid->root_objectid = BTRFS_I(inode)->root->objectid;
 	fid->gen = inode->i_generation;
 
@@ -178,13 +178,13 @@ static struct dentry *btrfs_get_parent(struct dentry *child)
 	if (!path)
 		return ERR_PTR(-ENOMEM);
 
-	if (dir->i_ino == BTRFS_FIRST_FREE_OBJECTID) {
+	if (btrfs_ino(dir) == BTRFS_FIRST_FREE_OBJECTID) {
 		key.objectid = root->root_key.objectid;
 		key.type = BTRFS_ROOT_BACKREF_KEY;
 		key.offset = (u64)-1;
 		root = root->fs_info->tree_root;
 	} else {
-		key.objectid = dir->i_ino;
+		key.objectid = btrfs_ino(dir);
 		key.type = BTRFS_INODE_REF_KEY;
 		key.offset = (u64)-1;
 	}
@@ -244,6 +244,7 @@ static int btrfs_get_name(struct dentry *parent, char *name,
 	struct btrfs_key key;
 	int name_len;
 	int ret;
+	u64 ino;
 
 	if (!dir || !inode)
 		return -EINVAL;
@@ -251,19 +252,21 @@ static int btrfs_get_name(struct dentry *parent, char *name,
 	if (!S_ISDIR(dir->i_mode))
 		return -EINVAL;
 
+	ino = btrfs_ino(inode);
+
 	path = btrfs_alloc_path();
 	if (!path)
 		return -ENOMEM;
 	path->leave_spinning = 1;
 
-	if (inode->i_ino == BTRFS_FIRST_FREE_OBJECTID) {
+	if (ino == BTRFS_FIRST_FREE_OBJECTID) {
 		key.objectid = BTRFS_I(inode)->root->root_key.objectid;
 		key.type = BTRFS_ROOT_BACKREF_KEY;
 		key.offset = (u64)-1;
 		root = root->fs_info->tree_root;
 	} else {
-		key.objectid = inode->i_ino;
-		key.offset = dir->i_ino;
+		key.objectid = ino;
+		key.offset = btrfs_ino(dir);
 		key.type = BTRFS_INODE_REF_KEY;
 	}
 
@@ -272,7 +275,7 @@ static int btrfs_get_name(struct dentry *parent, char *name,
 		btrfs_free_path(path);
 		return ret;
 	} else if (ret > 0) {
-		if (inode->i_ino == BTRFS_FIRST_FREE_OBJECTID) {
+		if (ino == BTRFS_FIRST_FREE_OBJECTID) {
 			path->slots[0]--;
 		} else {
 			btrfs_free_path(path);
@@ -281,11 +284,11 @@ static int btrfs_get_name(struct dentry *parent, char *name,
 	}
 	leaf = path->nodes[0];
 
-	if (inode->i_ino == BTRFS_FIRST_FREE_OBJECTID) {
-	       rref = btrfs_item_ptr(leaf, path->slots[0],
+	if (ino == BTRFS_FIRST_FREE_OBJECTID) {
+		rref = btrfs_item_ptr(leaf, path->slots[0],
 				     struct btrfs_root_ref);
-	       name_ptr = (unsigned long)(rref + 1);
-	       name_len = btrfs_root_ref_name_len(leaf, rref);
+		name_ptr = (unsigned long)(rref + 1);
+		name_len = btrfs_root_ref_name_len(leaf, rref);
 	} else {
 		iref = btrfs_item_ptr(leaf, path->slots[0],
 				      struct btrfs_inode_ref);

fs/btrfs/extent-tree.c

@@ -105,6 +105,7 @@ void btrfs_put_block_group(struct btrfs_block_group_cache *cache)
 		WARN_ON(cache->pinned > 0);
 		WARN_ON(cache->reserved > 0);
 		WARN_ON(cache->reserved_pinned > 0);
+		kfree(cache->free_space_ctl);
 		kfree(cache);
 	}
 }
@@ -3144,7 +3145,8 @@ int btrfs_check_data_free_space(struct inode *inode, u64 bytes)
 	/* make sure bytes are sectorsize aligned */
 	bytes = (bytes + root->sectorsize - 1) & ~((u64)root->sectorsize - 1);
 
-	if (root == root->fs_info->tree_root) {
+	if (root == root->fs_info->tree_root ||
+	    BTRFS_I(inode)->location.objectid == BTRFS_FREE_INO_OBJECTID) {
 		alloc_chunk = 0;
 		committed = 1;
 	}
@@ -4893,7 +4895,7 @@ wait_block_group_cache_progress(struct btrfs_block_group_cache *cache,
 		return 0;
 
 	wait_event(caching_ctl->wait, block_group_cache_done(cache) ||
-		   (cache->free_space >= num_bytes));
+		   (cache->free_space_ctl->free_space >= num_bytes));
 
 	put_caching_control(caching_ctl);
 	return 0;
@@ -7008,8 +7010,8 @@ static noinline int get_new_locations(struct inode *reloc_inode,
 
 	cur_pos = extent_key->objectid - offset;
 	last_byte = extent_key->objectid + extent_key->offset;
-	ret = btrfs_lookup_file_extent(NULL, root, path, reloc_inode->i_ino,
-				       cur_pos, 0);
+	ret = btrfs_lookup_file_extent(NULL, root, path,
+				       btrfs_ino(reloc_inode), cur_pos, 0);
 	if (ret < 0)
 		goto out;
 	if (ret > 0) {
@@ -7032,7 +7034,7 @@ static noinline int get_new_locations(struct inode *reloc_inode,
 		btrfs_item_key_to_cpu(leaf, &found_key, path->slots[0]);
 		if (found_key.offset != cur_pos ||
 		    found_key.type != BTRFS_EXTENT_DATA_KEY ||
-		    found_key.objectid != reloc_inode->i_ino)
+		    found_key.objectid != btrfs_ino(reloc_inode))
 			break;
 
 		fi = btrfs_item_ptr(leaf, path->slots[0],
@@ -7178,7 +7180,7 @@ next:
 				break;
 		}
 
-		if (inode && key.objectid != inode->i_ino) {
+		if (inode && key.objectid != btrfs_ino(inode)) {
 			BUG_ON(extent_locked);
 			btrfs_release_path(root, path);
 			mutex_unlock(&inode->i_mutex);
@@ -7487,7 +7489,7 @@ static noinline int invalidate_extent_cache(struct btrfs_root *root,
 			continue;
 		if (btrfs_file_extent_disk_bytenr(leaf, fi) == 0)
 			continue;
-		if (!inode || inode->i_ino != key.objectid) {
+		if (!inode || btrfs_ino(inode) != key.objectid) {
 			iput(inode);
 			inode = btrfs_ilookup(target_root->fs_info->sb,
 					      key.objectid, target_root, 1);
@@ -8555,10 +8557,16 @@ int btrfs_read_block_groups(struct btrfs_root *root)
 			ret = -ENOMEM;
 			goto error;
 		}
+		cache->free_space_ctl = kzalloc(sizeof(*cache->free_space_ctl),
+						GFP_NOFS);
+		if (!cache->free_space_ctl) {
+			kfree(cache);
+			ret = -ENOMEM;
+			goto error;
+		}
 
 		atomic_set(&cache->count, 1);
 		spin_lock_init(&cache->lock);
-		spin_lock_init(&cache->tree_lock);
 		cache->fs_info = info;
 		INIT_LIST_HEAD(&cache->list);
 		INIT_LIST_HEAD(&cache->cluster_list);
@@ -8566,14 +8574,6 @@ int btrfs_read_block_groups(struct btrfs_root *root)
 		if (need_clear)
 			cache->disk_cache_state = BTRFS_DC_CLEAR;
 
-		/*
-		 * we only want to have 32k of ram per block group for keeping
-		 * track of free space, and if we pass 1/2 of that we want to
-		 * start converting things over to using bitmaps
-		 */
-		cache->extents_thresh = ((1024 * 32) / 2) /
-			sizeof(struct btrfs_free_space);
-
 		read_extent_buffer(leaf, &cache->item,
 				   btrfs_item_ptr_offset(leaf, path->slots[0]),
 				   sizeof(cache->item));
@@ -8584,6 +8584,8 @@ int btrfs_read_block_groups(struct btrfs_root *root)
 		cache->flags = btrfs_block_group_flags(&cache->item);
 		cache->sectorsize = root->sectorsize;
 
+		btrfs_init_free_space_ctl(cache);
+
 		/*
 		 * We need to exclude the super stripes now so that the space
 		 * info has super bytes accounted for, otherwise we'll think
@@ -8670,6 +8672,12 @@ int btrfs_make_block_group(struct btrfs_trans_handle *trans,
 	cache = kzalloc(sizeof(*cache), GFP_NOFS);
 	if (!cache)
 		return -ENOMEM;
+	cache->free_space_ctl = kzalloc(sizeof(*cache->free_space_ctl),
+					GFP_NOFS);
+	if (!cache->free_space_ctl) {
+		kfree(cache);
+		return -ENOMEM;
+	}
 
 	cache->key.objectid = chunk_offset;
 	cache->key.offset = size;
@@ -8677,19 +8685,13 @@ int btrfs_make_block_group(struct btrfs_trans_handle *trans,
 	cache->sectorsize = root->sectorsize;
 	cache->fs_info = root->fs_info;
 
-	/*
-	 * we only want to have 32k of ram per block group for keeping track
-	 * of free space, and if we pass 1/2 of that we want to start
-	 * converting things over to using bitmaps
-	 */
-	cache->extents_thresh = ((1024 * 32) / 2) /
-		sizeof(struct btrfs_free_space);
 	atomic_set(&cache->count, 1);
 	spin_lock_init(&cache->lock);
-	spin_lock_init(&cache->tree_lock);
 	INIT_LIST_HEAD(&cache->list);
 	INIT_LIST_HEAD(&cache->cluster_list);
 
+	btrfs_init_free_space_ctl(cache);
+
 	btrfs_set_block_group_used(&cache->item, bytes_used);
 	btrfs_set_block_group_chunk_objectid(&cache->item, chunk_objectid);
 	cache->flags = type;

fs/btrfs/extent_io.c

@@ -3030,7 +3030,7 @@ int extent_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo,
 	 * because there might be preallocation past i_size
 	 */
 	ret = btrfs_lookup_file_extent(NULL, BTRFS_I(inode)->root,
-				       path, inode->i_ino, -1, 0);
+				       path, btrfs_ino(inode), -1, 0);
 	if (ret < 0) {
 		btrfs_free_path(path);
 		return ret;
@@ -3043,7 +3043,7 @@ int extent_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo,
 	found_type = btrfs_key_type(&found_key);
 
 	/* No extents, but there might be delalloc bits */
-	if (found_key.objectid != inode->i_ino ||
+	if (found_key.objectid != btrfs_ino(inode) ||
 	    found_type != BTRFS_EXTENT_DATA_KEY) {
 		/* have to trust i_size as the end */
 		last = (u64)-1;

fs/btrfs/file-item.c

@@ -208,8 +208,9 @@ static int __btrfs_lookup_bio_sums(struct btrfs_root *root,
 						EXTENT_NODATASUM, GFP_NOFS);
 				} else {
 					printk(KERN_INFO "btrfs no csum found "
-					       "for inode %lu start %llu\n",
-					       inode->i_ino,
+					       "for inode %llu start %llu\n",
+					       (unsigned long long)
+					       btrfs_ino(inode),
 					       (unsigned long long)offset);
 				}
 				item = NULL;

fs/btrfs/file.c

@@ -298,6 +298,7 @@ int btrfs_drop_extents(struct btrfs_trans_handle *trans, struct inode *inode,
 	struct btrfs_path *path;
 	struct btrfs_key key;
 	struct btrfs_key new_key;
+	u64 ino = btrfs_ino(inode);
 	u64 search_start = start;
 	u64 disk_bytenr = 0;
 	u64 num_bytes = 0;
@@ -318,14 +319,14 @@ int btrfs_drop_extents(struct btrfs_trans_handle *trans, struct inode *inode,
 
 	while (1) {
 		recow = 0;
-		ret = btrfs_lookup_file_extent(trans, root, path, inode->i_ino,
+		ret = btrfs_lookup_file_extent(trans, root, path, ino,
 					       search_start, -1);
 		if (ret < 0)
 			break;
 		if (ret > 0 && path->slots[0] > 0 && search_start == start) {
 			leaf = path->nodes[0];
 			btrfs_item_key_to_cpu(leaf, &key, path->slots[0] - 1);
-			if (key.objectid == inode->i_ino &&
+			if (key.objectid == ino &&
 			    key.type == BTRFS_EXTENT_DATA_KEY)
 				path->slots[0]--;
 		}
@@ -346,7 +347,7 @@ next_slot:
 		}
 
 		btrfs_item_key_to_cpu(leaf, &key, path->slots[0]);
-		if (key.objectid > inode->i_ino ||
+		if (key.objectid > ino ||
 		    key.type > BTRFS_EXTENT_DATA_KEY || key.offset >= end)
 			break;
 
@@ -592,6 +593,7 @@ int btrfs_mark_extent_written(struct btrfs_trans_handle *trans,
 	int del_slot = 0;
 	int recow;
 	int ret;
+	u64 ino = btrfs_ino(inode);
 
 	btrfs_drop_extent_cache(inode, start, end - 1, 0);
 
@@ -600,7 +602,7 @@ int btrfs_mark_extent_written(struct btrfs_trans_handle *trans,
 again:
 	recow = 0;
 	split = start;
-	key.objectid = inode->i_ino;
+	key.objectid = ino;
 	key.type = BTRFS_EXTENT_DATA_KEY;
 	key.offset = split;
 
@@ -612,8 +614,7 @@ again:
 
 	leaf = path->nodes[0];
 	btrfs_item_key_to_cpu(leaf, &key, path->slots[0]);
-	BUG_ON(key.objectid != inode->i_ino ||
-	       key.type != BTRFS_EXTENT_DATA_KEY);
+	BUG_ON(key.objectid != ino || key.type != BTRFS_EXTENT_DATA_KEY);
 	fi = btrfs_item_ptr(leaf, path->slots[0],
 			    struct btrfs_file_extent_item);
 	BUG_ON(btrfs_file_extent_type(leaf, fi) !=
@@ -630,7 +631,7 @@ again:
 		other_start = 0;
 		other_end = start;
 		if (extent_mergeable(leaf, path->slots[0] - 1,
-				     inode->i_ino, bytenr, orig_offset,
+				     ino, bytenr, orig_offset,
 				     &other_start, &other_end)) {
 			new_key.offset = end;
 			btrfs_set_item_key_safe(trans, root, path, &new_key);
@@ -653,7 +654,7 @@ again:
 		other_start = end;
 		other_end = 0;
 		if (extent_mergeable(leaf, path->slots[0] + 1,
-				     inode->i_ino, bytenr, orig_offset,
+				     ino, bytenr, orig_offset,
 				     &other_start, &other_end)) {
 			fi = btrfs_item_ptr(leaf, path->slots[0],
 					    struct btrfs_file_extent_item);
@@ -702,7 +703,7 @@ again:
 
 		ret = btrfs_inc_extent_ref(trans, root, bytenr, num_bytes, 0,
 					   root->root_key.objectid,
-					   inode->i_ino, orig_offset);
+					   ino, orig_offset);
 		BUG_ON(ret);
 
 		if (split == start) {
@@ -718,7 +719,7 @@ again:
 	other_start = end;
 	other_end = 0;
 	if (extent_mergeable(leaf, path->slots[0] + 1,
-			     inode->i_ino, bytenr, orig_offset,
+			     ino, bytenr, orig_offset,
 			     &other_start, &other_end)) {
 		if (recow) {
 			btrfs_release_path(root, path);
@@ -729,13 +730,13 @@ again:
 		del_nr++;
 		ret = btrfs_free_extent(trans, root, bytenr, num_bytes,
 					0, root->root_key.objectid,
-					inode->i_ino, orig_offset);
+					ino, orig_offset);
 		BUG_ON(ret);
 	}
 	other_start = 0;
 	other_end = start;
 	if (extent_mergeable(leaf, path->slots[0] - 1,
-			     inode->i_ino, bytenr, orig_offset,
+			     ino, bytenr, orig_offset,
 			     &other_start, &other_end)) {
 		if (recow) {
 			btrfs_release_path(root, path);
@@ -746,7 +747,7 @@ again:
 		del_nr++;
 		ret = btrfs_free_extent(trans, root, bytenr, num_bytes,
 					0, root->root_key.objectid,
-					inode->i_ino, orig_offset);
+					ino, orig_offset);
 		BUG_ON(ret);
 	}
 	if (del_nr == 0) {

fs/btrfs/free-space-cache.c

@@ -25,18 +25,17 @@
 #include "transaction.h"
 #include "disk-io.h"
 #include "extent_io.h"
+#include "inode-map.h"
 
 #define BITS_PER_BITMAP		(PAGE_CACHE_SIZE * 8)
 #define MAX_CACHE_BYTES_PER_GIG	(32 * 1024)
 
-static void recalculate_thresholds(struct btrfs_block_group_cache
-				   *block_group);
-static int link_free_space(struct btrfs_block_group_cache *block_group,
+static int link_free_space(struct btrfs_free_space_ctl *ctl,
 			   struct btrfs_free_space *info);
 
-struct inode *lookup_free_space_inode(struct btrfs_root *root,
-				      struct btrfs_block_group_cache
-				      *block_group, struct btrfs_path *path)
+static struct inode *__lookup_free_space_inode(struct btrfs_root *root,
+					       struct btrfs_path *path,
+					       u64 offset)
 {
 	struct btrfs_key key;
 	struct btrfs_key location;
@@ -46,15 +45,8 @@ struct inode *lookup_free_space_inode(struct btrfs_root *root,
 	struct inode *inode = NULL;
 	int ret;
 
-	spin_lock(&block_group->lock);
-	if (block_group->inode)
-		inode = igrab(block_group->inode);
-	spin_unlock(&block_group->lock);
-	if (inode)
-		return inode;
-
 	key.objectid = BTRFS_FREE_SPACE_OBJECTID;
-	key.offset = block_group->key.objectid;
+	key.offset = offset;
 	key.type = 0;
 
 	ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
@@ -84,6 +76,27 @@ struct inode *lookup_free_space_inode(struct btrfs_root *root,
 
 	inode->i_mapping->flags &= ~__GFP_FS;
 
+	return inode;
+}
+
+struct inode *lookup_free_space_inode(struct btrfs_root *root,
+				      struct btrfs_block_group_cache
+				      *block_group, struct btrfs_path *path)
+{
+	struct inode *inode = NULL;
+
+	spin_lock(&block_group->lock);
+	if (block_group->inode)
+		inode = igrab(block_group->inode);
+	spin_unlock(&block_group->lock);
+	if (inode)
+		return inode;
+
+	inode = __lookup_free_space_inode(root, path,
+					  block_group->key.objectid);
+	if (IS_ERR(inode))
+		return inode;
+
 	spin_lock(&block_group->lock);
 	if (!root->fs_info->closing) {
 		block_group->inode = igrab(inode);
@@ -94,24 +107,18 @@ struct inode *lookup_free_space_inode(struct btrfs_root *root,
 	return inode;
 }
 
-int create_free_space_inode(struct btrfs_root *root,
-			    struct btrfs_trans_handle *trans,
-			    struct btrfs_block_group_cache *block_group,
-			    struct btrfs_path *path)
+int __create_free_space_inode(struct btrfs_root *root,
+			      struct btrfs_trans_handle *trans,
+			      struct btrfs_path *path, u64 ino, u64 offset)
 {
 	struct btrfs_key key;
 	struct btrfs_disk_key disk_key;
 	struct btrfs_free_space_header *header;
 	struct btrfs_inode_item *inode_item;
 	struct extent_buffer *leaf;
-	u64 objectid;
 	int ret;
 
-	ret = btrfs_find_free_objectid(trans, root, 0, &objectid);
-	if (ret < 0)
-		return ret;
-
-	ret = btrfs_insert_empty_inode(trans, root, path, objectid);
+	ret = btrfs_insert_empty_inode(trans, root, path, ino);
 	if (ret)
 		return ret;
 
@@ -131,13 +138,12 @@ int create_free_space_inode(struct btrfs_root *root,
 			      BTRFS_INODE_PREALLOC | BTRFS_INODE_NODATASUM);
 	btrfs_set_inode_nlink(leaf, inode_item, 1);
 	btrfs_set_inode_transid(leaf, inode_item, trans->transid);
-	btrfs_set_inode_block_group(leaf, inode_item,
-				    block_group->key.objectid);
+	btrfs_set_inode_block_group(leaf, inode_item, offset);
 	btrfs_mark_buffer_dirty(leaf);
 	btrfs_release_path(root, path);
 
 	key.objectid = BTRFS_FREE_SPACE_OBJECTID;
-	key.offset = block_group->key.objectid;
+	key.offset = offset;
 	key.type = 0;
 
 	ret = btrfs_insert_empty_item(trans, root, path, &key,
@@ -157,6 +163,22 @@ int create_free_space_inode(struct btrfs_root *root,
 	return 0;
 }
 
+int create_free_space_inode(struct btrfs_root *root,
+			    struct btrfs_trans_handle *trans,
+			    struct btrfs_block_group_cache *block_group,
+			    struct btrfs_path *path)
+{
+	int ret;
+	u64 ino;
+
+	ret = btrfs_find_free_objectid(root, &ino);
+	if (ret < 0)
+		return ret;
+
+	return __create_free_space_inode(root, trans, path, ino,
+					 block_group->key.objectid);
+}
+
 int btrfs_truncate_free_space_cache(struct btrfs_root *root,
 				    struct btrfs_trans_handle *trans,
 				    struct btrfs_path *path,
@@ -187,7 +209,8 @@ int btrfs_truncate_free_space_cache(struct btrfs_root *root,
 		return ret;
 	}
 
-	return btrfs_update_inode(trans, root, inode);
+	ret = btrfs_update_inode(trans, root, inode);
+	return ret;
 }
 
 static int readahead_cache(struct inode *inode)
@@ -209,15 +232,13 @@ static int readahead_cache(struct inode *inode)
 	return 0;
 }
 
-int load_free_space_cache(struct btrfs_fs_info *fs_info,
-			  struct btrfs_block_group_cache *block_group)
+int __load_free_space_cache(struct btrfs_root *root, struct inode *inode,
+			    struct btrfs_free_space_ctl *ctl,
+			    struct btrfs_path *path, u64 offset)
 {
-	struct btrfs_root *root = fs_info->tree_root;
-	struct inode *inode;
 	struct btrfs_free_space_header *header;
 	struct extent_buffer *leaf;
 	struct page *page;
-	struct btrfs_path *path;
 	u32 *checksums = NULL, *crc;
 	char *disk_crcs = NULL;
 	struct btrfs_key key;
@@ -225,76 +246,47 @@ int load_free_space_cache(struct btrfs_fs_info *fs_info,
 	u64 num_entries;
 	u64 num_bitmaps;
 	u64 generation;
-	u64 used = btrfs_block_group_used(&block_group->item);
 	u32 cur_crc = ~(u32)0;
 	pgoff_t index = 0;
 	unsigned long first_page_offset;
 	int num_checksums;
-	int ret = 0;
-
-	/*
-	 * If we're unmounting then just return, since this does a search on the
-	 * normal root and not the commit root and we could deadlock.
-	 */
-	smp_mb();
-	if (fs_info->closing)
-		return 0;
-
-	/*
-	 * If this block group has been marked to be cleared for one reason or
-	 * another then we can't trust the on disk cache, so just return.
-	 */
-	spin_lock(&block_group->lock);
-	if (block_group->disk_cache_state != BTRFS_DC_WRITTEN) {
-		spin_unlock(&block_group->lock);
-		return 0;
-	}
-	spin_unlock(&block_group->lock);
+	int ret = 0, ret2;
 
 	INIT_LIST_HEAD(&bitmaps);
 
-	path = btrfs_alloc_path();
-	if (!path)
-		return 0;
-
-	inode = lookup_free_space_inode(root, block_group, path);
-	if (IS_ERR(inode)) {
-		btrfs_free_path(path);
-		return 0;
-	}
-
 	/* Nothing in the space cache, goodbye */
-	if (!i_size_read(inode)) {
-		btrfs_free_path(path);
+	if (!i_size_read(inode))
 		goto out;
-	}
 
 	key.objectid = BTRFS_FREE_SPACE_OBJECTID;
-	key.offset = block_group->key.objectid;
+	key.offset = offset;
 	key.type = 0;
 
 	ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
-	if (ret) {
-		btrfs_free_path(path);
+	if (ret < 0)
+		goto out;
+	else if (ret > 0) {
+		btrfs_release_path(root, path);
+		ret = 0;
 		goto out;
 	}
 
+	ret = -1;
+
 	leaf = path->nodes[0];
 	header = btrfs_item_ptr(leaf, path->slots[0],
 				struct btrfs_free_space_header);
 	num_entries = btrfs_free_space_entries(leaf, header);
 	num_bitmaps = btrfs_free_space_bitmaps(leaf, header);
 	generation = btrfs_free_space_generation(leaf, header);
-	btrfs_free_path(path);
+	btrfs_release_path(root, path);
 
 	if (BTRFS_I(inode)->generation != generation) {
 		printk(KERN_ERR "btrfs: free space inode generation (%llu) did"
-		       " not match free space cache generation (%llu) for "
-		       "block group %llu\n",
+		       " not match free space cache generation (%llu)\n",
 		       (unsigned long long)BTRFS_I(inode)->generation,
-		       (unsigned long long)generation,
-		       (unsigned long long)block_group->key.objectid);
-		goto free_cache;
+		       (unsigned long long)generation);
+		goto out;
 	}
 
 	if (!num_entries)
@@ -311,10 +303,8 @@ int load_free_space_cache(struct btrfs_fs_info *fs_info,
 		goto out;
 
 	ret = readahead_cache(inode);
-	if (ret) {
-		ret = 0;
+	if (ret)
 		goto out;
-	}
 
 	while (1) {
 		struct btrfs_free_space_entry *entry;
@@ -333,10 +323,8 @@ int load_free_space_cache(struct btrfs_fs_info *fs_info,
 		}
 
 		page = grab_cache_page(inode->i_mapping, index);
-		if (!page) {
-			ret = 0;
+		if (!page)
 			goto free_cache;
-		}
 
 		if (!PageUptodate(page)) {
 			btrfs_readpage(NULL, page);
@@ -345,9 +333,7 @@ int load_free_space_cache(struct btrfs_fs_info *fs_info,
 				unlock_page(page);
 				page_cache_release(page);
 				printk(KERN_ERR "btrfs: error reading free "
-				       "space cache: %llu\n",
-				       (unsigned long long)
-				       block_group->key.objectid);
+				       "space cache\n");
 				goto free_cache;
 			}
 		}
@@ -360,13 +346,10 @@ int load_free_space_cache(struct btrfs_fs_info *fs_info,
 			gen = addr + (sizeof(u32) * num_checksums);
 			if (*gen != BTRFS_I(inode)->generation) {
 				printk(KERN_ERR "btrfs: space cache generation"
-				       " (%llu) does not match inode (%llu) "
-				       "for block group %llu\n",
+				       " (%llu) does not match inode (%llu)\n",
 				       (unsigned long long)*gen,
 				       (unsigned long long)
-				       BTRFS_I(inode)->generation,
-				       (unsigned long long)
-				       block_group->key.objectid);
+				       BTRFS_I(inode)->generation);
 				kunmap(page);
 				unlock_page(page);
 				page_cache_release(page);
@@ -382,9 +365,8 @@ int load_free_space_cache(struct btrfs_fs_info *fs_info,
 					  PAGE_CACHE_SIZE - start_offset);
 		btrfs_csum_final(cur_crc, (char *)&cur_crc);
 		if (cur_crc != *crc) {
-			printk(KERN_ERR "btrfs: crc mismatch for page %lu in "
-			       "block group %llu\n", index,
-			       (unsigned long long)block_group->key.objectid);
+			printk(KERN_ERR "btrfs: crc mismatch for page %lu\n",
+			       index);
 			kunmap(page);
 			unlock_page(page);
 			page_cache_release(page);
@@ -417,9 +399,9 @@ int load_free_space_cache(struct btrfs_fs_info *fs_info,
 			}
 
 			if (entry->type == BTRFS_FREE_SPACE_EXTENT) {
-				spin_lock(&block_group->tree_lock);
-				ret = link_free_space(block_group, e);
-				spin_unlock(&block_group->tree_lock);
+				spin_lock(&ctl->tree_lock);
+				ret = link_free_space(ctl, e);
+				spin_unlock(&ctl->tree_lock);
 				BUG_ON(ret);
 			} else {
 				e->bitmap = kzalloc(PAGE_CACHE_SIZE, GFP_NOFS);
@@ -431,11 +413,11 @@ int load_free_space_cache(struct btrfs_fs_info *fs_info,
 					page_cache_release(page);
 					goto free_cache;
 				}
-				spin_lock(&block_group->tree_lock);
-				ret = link_free_space(block_group, e);
-				block_group->total_bitmaps++;
-				recalculate_thresholds(block_group);
-				spin_unlock(&block_group->tree_lock);
+				spin_lock(&ctl->tree_lock);
+				ret2 = link_free_space(ctl, e);
+				ctl->total_bitmaps++;
+				ctl->op->recalc_thresholds(ctl);
+				spin_unlock(&ctl->tree_lock);
 				list_add_tail(&e->list, &bitmaps);
 			}
 
@@ -471,41 +453,97 @@ next:
 		index++;
 	}
 
-	spin_lock(&block_group->tree_lock);
-	if (block_group->free_space != (block_group->key.offset - used -
-					block_group->bytes_super)) {
-		spin_unlock(&block_group->tree_lock);
-		printk(KERN_ERR "block group %llu has an wrong amount of free "
-		       "space\n", block_group->key.objectid);
-		ret = 0;
-		goto free_cache;
-	}
-	spin_unlock(&block_group->tree_lock);
-
 	ret = 1;
 out:
 	kfree(checksums);
 	kfree(disk_crcs);
-	iput(inode);
 	return ret;
-
 free_cache:
-	/* This cache is bogus, make sure it gets cleared */
+	__btrfs_remove_free_space_cache(ctl);
+	goto out;
+}
+
+int load_free_space_cache(struct btrfs_fs_info *fs_info,
+			  struct btrfs_block_group_cache *block_group)
+{
+	struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl;
+	struct btrfs_root *root = fs_info->tree_root;
+	struct inode *inode;
+	struct btrfs_path *path;
+	int ret;
+	bool matched;
+	u64 used = btrfs_block_group_used(&block_group->item);
+
+	/*
+	 * If we're unmounting then just return, since this does a search on the
+	 * normal root and not the commit root and we could deadlock.
+	 */
+	smp_mb();
+	if (fs_info->closing)
+		return 0;
+
+	/*
+	 * If this block group has been marked to be cleared for one reason or
+	 * another then we can't trust the on disk cache, so just return.
+	 */
 	spin_lock(&block_group->lock);
-	block_group->disk_cache_state = BTRFS_DC_CLEAR;
+	if (block_group->disk_cache_state != BTRFS_DC_WRITTEN) {
+		spin_unlock(&block_group->lock);
+		return 0;
+	}
 	spin_unlock(&block_group->lock);
-	btrfs_remove_free_space_cache(block_group);
-	goto out;
+
+	path = btrfs_alloc_path();
+	if (!path)
+		return 0;
+
+	inode = lookup_free_space_inode(root, block_group, path);
+	if (IS_ERR(inode)) {
+		btrfs_free_path(path);
+		return 0;
+	}
+
+	ret = __load_free_space_cache(fs_info->tree_root, inode, ctl,
+				      path, block_group->key.objectid);
+	btrfs_free_path(path);
+	if (ret <= 0)
+		goto out;
+
+	spin_lock(&ctl->tree_lock);
+	matched = (ctl->free_space == (block_group->key.offset - used -
+				       block_group->bytes_super));
+	spin_unlock(&ctl->tree_lock);
+
+	if (!matched) {
+		__btrfs_remove_free_space_cache(ctl);
+		printk(KERN_ERR "block group %llu has an wrong amount of free "
+		       "space\n", block_group->key.objectid);
+		ret = -1;
+	}
+out:
+	if (ret < 0) {
+		/* This cache is bogus, make sure it gets cleared */
+		spin_lock(&block_group->lock);
+		block_group->disk_cache_state = BTRFS_DC_CLEAR;
+		spin_unlock(&block_group->lock);
+		ret = 0;
+
+		printk(KERN_ERR "btrfs: failed to load free space cache "
+		       "for block group %llu\n", block_group->key.objectid);
+	}
+
+	iput(inode);
+	return ret;
 }
 
-int btrfs_write_out_cache(struct btrfs_root *root,
-			  struct btrfs_trans_handle *trans,
-			  struct btrfs_block_group_cache *block_group,
-			  struct btrfs_path *path)
+int __btrfs_write_out_cache(struct btrfs_root *root, struct inode *inode,
+			    struct btrfs_free_space_ctl *ctl,
+			    struct btrfs_block_group_cache *block_group,
+			    struct btrfs_trans_handle *trans,
+			    struct btrfs_path *path, u64 offset)
 {
 	struct btrfs_free_space_header *header;
 	struct extent_buffer *leaf;
-	struct inode *inode;
 	struct rb_node *node;
 	struct list_head *pos, *n;
 	struct page **pages;
@@ -522,35 +560,18 @@ int btrfs_write_out_cache(struct btrfs_root *root,
 	int index = 0, num_pages = 0;
 	int entries = 0;
 	int bitmaps = 0;
-	int ret = 0;
+	int ret = -1;
 	bool next_page = false;
 	bool out_of_space = false;
 
-	root = root->fs_info->tree_root;
-
 	INIT_LIST_HEAD(&bitmap_list);
 
-	spin_lock(&block_group->lock);
-	if (block_group->disk_cache_state < BTRFS_DC_SETUP) {
-		spin_unlock(&block_group->lock);
-		return 0;
-	}
-	spin_unlock(&block_group->lock);
-
-	inode = lookup_free_space_inode(root, block_group, path);
-	if (IS_ERR(inode))
+	node = rb_first(&ctl->free_space_offset);
+	if (!node)
 		return 0;
 
-	if (!i_size_read(inode)) {
-		iput(inode);
-		return 0;
-	}
-
-	node = rb_first(&block_group->free_space_offset);
-	if (!node) {
-		iput(inode);
-		return 0;
-	}
+	if (!i_size_read(inode))
+		return -1;
 
 	num_pages = (i_size_read(inode) + PAGE_CACHE_SIZE - 1) >>
 		PAGE_CACHE_SHIFT;
@@ -560,16 +581,13 @@ int btrfs_write_out_cache(struct btrfs_root *root,
 
 	/* We need a checksum per page. */
 	crc = checksums = kzalloc(sizeof(u32) * num_pages, GFP_NOFS);
-	if (!crc) {
-		iput(inode);
-		return 0;
-	}
+	if (!crc)
+		return -1;
 
 	pages = kzalloc(sizeof(struct page *) * num_pages, GFP_NOFS);
 	if (!pages) {
 		kfree(crc);
-		iput(inode);
-		return 0;
+		return -1;
 	}
 
 	/* Since the first page has all of our checksums and our generation we
@@ -579,7 +597,7 @@ int btrfs_write_out_cache(struct btrfs_root *root,
 	first_page_offset = (sizeof(u32) * num_pages) + sizeof(u64);
 
 	/* Get the cluster for this block_group if it exists */
-	if (!list_empty(&block_group->cluster_list))
+	if (block_group && !list_empty(&block_group->cluster_list))
 		cluster = list_entry(block_group->cluster_list.next,
 				     struct btrfs_free_cluster,
 				     block_group_list);
@@ -621,7 +639,8 @@ int btrfs_write_out_cache(struct btrfs_root *root,
 	 * When searching for pinned extents, we need to start at our start
 	 * offset.
 	 */
-	start = block_group->key.objectid;
+	if (block_group)
+		start = block_group->key.objectid;
 
 	/* Write out the extent entries */
 	do {
@@ -679,8 +698,9 @@ int btrfs_write_out_cache(struct btrfs_root *root,
 		 * We want to add any pinned extents to our free space cache
 		 * so we don't leak the space
 		 */
-		while (!next_page && (start < block_group->key.objectid +
-				      block_group->key.offset)) {
+		while (block_group && !next_page &&
+		       (start < block_group->key.objectid +
+			block_group->key.offset)) {
 			ret = find_first_extent_bit(unpin, start, &start, &end,
 						    EXTENT_DIRTY);
 			if (ret) {
@@ -798,12 +818,12 @@ int btrfs_write_out_cache(struct btrfs_root *root,
 	filemap_write_and_wait(inode->i_mapping);
 
 	key.objectid = BTRFS_FREE_SPACE_OBJECTID;
-	key.offset = block_group->key.objectid;
+	key.offset = offset;
 	key.type = 0;
 
 	ret = btrfs_search_slot(trans, root, &key, path, 1, 1);
 	if (ret < 0) {
-		ret = 0;
+		ret = -1;
 		clear_extent_bit(&BTRFS_I(inode)->io_tree, 0, bytes - 1,
 				 EXTENT_DIRTY | EXTENT_DELALLOC |
 				 EXTENT_DO_ACCOUNTING, 0, 0, NULL, GFP_NOFS);
@@ -816,8 +836,8 @@ int btrfs_write_out_cache(struct btrfs_root *root,
 		path->slots[0]--;
 		btrfs_item_key_to_cpu(leaf, &found_key, path->slots[0]);
 		if (found_key.objectid != BTRFS_FREE_SPACE_OBJECTID ||
-		    found_key.offset != block_group->key.objectid) {
-			ret = 0;
+		    found_key.offset != offset) {
+			ret = -1;
 			clear_extent_bit(&BTRFS_I(inode)->io_tree, 0, bytes - 1,
 					 EXTENT_DIRTY | EXTENT_DELALLOC |
 					 EXTENT_DO_ACCOUNTING, 0, 0, NULL,
@@ -837,44 +857,78 @@ int btrfs_write_out_cache(struct btrfs_root *root,
 	ret = 1;
 
 out_free:
-	if (ret == 0) {
+	if (ret != 1) {
 		invalidate_inode_pages2_range(inode->i_mapping, 0, index);
-		spin_lock(&block_group->lock);
-		block_group->disk_cache_state = BTRFS_DC_ERROR;
-		spin_unlock(&block_group->lock);
 		BTRFS_I(inode)->generation = 0;
 	}
 	kfree(checksums);
 	kfree(pages);
 	btrfs_update_inode(trans, root, inode);
+	return ret;
+}
+
+int btrfs_write_out_cache(struct btrfs_root *root,
+			  struct btrfs_trans_handle *trans,
+			  struct btrfs_block_group_cache *block_group,
+			  struct btrfs_path *path)
+{
+	struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl;
+	struct inode *inode;
+	int ret = 0;
+
+	root = root->fs_info->tree_root;
+
+	spin_lock(&block_group->lock);
+	if (block_group->disk_cache_state < BTRFS_DC_SETUP) {
+		spin_unlock(&block_group->lock);
+		return 0;
+	}
+	spin_unlock(&block_group->lock);
+
+	inode = lookup_free_space_inode(root, block_group, path);
+	if (IS_ERR(inode))
+		return 0;
+
+	ret = __btrfs_write_out_cache(root, inode, ctl, block_group, trans,
+				      path, block_group->key.objectid);
+	if (ret < 0) {
+		spin_lock(&block_group->lock);
+		block_group->disk_cache_state = BTRFS_DC_ERROR;
+		spin_unlock(&block_group->lock);
+		ret = 0;
+
+		printk(KERN_ERR "btrfs: failed to write free space cace "
+		       "for block group %llu\n", block_group->key.objectid);
+	}
+
 	iput(inode);
 	return ret;
 }
 
-static inline unsigned long offset_to_bit(u64 bitmap_start, u64 sectorsize,
+static inline unsigned long offset_to_bit(u64 bitmap_start, u32 unit,
 					  u64 offset)
 {
 	BUG_ON(offset < bitmap_start);
 	offset -= bitmap_start;
-	return (unsigned long)(div64_u64(offset, sectorsize));
+	return (unsigned long)(div_u64(offset, unit));
 }
 
-static inline unsigned long bytes_to_bits(u64 bytes, u64 sectorsize)
+static inline unsigned long bytes_to_bits(u64 bytes, u32 unit)
 {
-	return (unsigned long)(div64_u64(bytes, sectorsize));
+	return (unsigned long)(div_u64(bytes, unit));
 }
 
-static inline u64 offset_to_bitmap(struct btrfs_block_group_cache *block_group,
+static inline u64 offset_to_bitmap(struct btrfs_free_space_ctl *ctl,
 				   u64 offset)
 {
 	u64 bitmap_start;
 	u64 bytes_per_bitmap;
 
-	bytes_per_bitmap = BITS_PER_BITMAP * block_group->sectorsize;
-	bitmap_start = offset - block_group->key.objectid;
+	bytes_per_bitmap = BITS_PER_BITMAP * ctl->unit;
+	bitmap_start = offset - ctl->start;
 	bitmap_start = div64_u64(bitmap_start, bytes_per_bitmap);
 	bitmap_start *= bytes_per_bitmap;
-	bitmap_start += block_group->key.objectid;
+	bitmap_start += ctl->start;
 
 	return bitmap_start;
 }
@@ -932,10 +986,10 @@ static int tree_insert_offset(struct rb_root *root, u64 offset,
  * offset.
  */
 static struct btrfs_free_space *
-tree_search_offset(struct btrfs_block_group_cache *block_group,
+tree_search_offset(struct btrfs_free_space_ctl *ctl,
 		   u64 offset, int bitmap_only, int fuzzy)
 {
-	struct rb_node *n = block_group->free_space_offset.rb_node;
+	struct rb_node *n = ctl->free_space_offset.rb_node;
 	struct btrfs_free_space *entry, *prev = NULL;
 
 	/* find entry that is closest to the 'offset' */
@@ -1031,8 +1085,7 @@ tree_search_offset(struct btrfs_block_group_cache *block_group,
 				break;
 			}
 		}
-		if (entry->offset + BITS_PER_BITMAP *
-		    block_group->sectorsize > offset)
+		if (entry->offset + BITS_PER_BITMAP * ctl->unit > offset)
 			return entry;
 	} else if (entry->offset + entry->bytes > offset)
 		return entry;
@@ -1043,7 +1096,7 @@ tree_search_offset(struct btrfs_block_group_cache *block_group,
 	while (1) {
 		if (entry->bitmap) {
 			if (entry->offset + BITS_PER_BITMAP *
-			    block_group->sectorsize > offset)
+			    ctl->unit > offset)
 				break;
 		} else {
 			if (entry->offset + entry->bytes > offset)
@@ -1059,42 +1112,47 @@ tree_search_offset(struct btrfs_block_group_cache *block_group,
 }
 
 static inline void
-__unlink_free_space(struct btrfs_block_group_cache *block_group,
+__unlink_free_space(struct btrfs_free_space_ctl *ctl,
 		    struct btrfs_free_space *info)
 {
-	rb_erase(&info->offset_index, &block_group->free_space_offset);
-	block_group->free_extents--;
+	rb_erase(&info->offset_index, &ctl->free_space_offset);
+	ctl->free_extents--;
 }
 
-static void unlink_free_space(struct btrfs_block_group_cache *block_group,
+static void unlink_free_space(struct btrfs_free_space_ctl *ctl,
 			      struct btrfs_free_space *info)
 {
-	__unlink_free_space(block_group, info);
-	block_group->free_space -= info->bytes;
+	__unlink_free_space(ctl, info);
+	ctl->free_space -= info->bytes;
 }
 
-static int link_free_space(struct btrfs_block_group_cache *block_group,
+static int link_free_space(struct btrfs_free_space_ctl *ctl,
 			   struct btrfs_free_space *info)
 {
 	int ret = 0;
 
 	BUG_ON(!info->bitmap && !info->bytes);
-	ret = tree_insert_offset(&block_group->free_space_offset, info->offset,
+	ret = tree_insert_offset(&ctl->free_space_offset, info->offset,
 				 &info->offset_index, (info->bitmap != NULL));
 	if (ret)
 		return ret;
 
-	block_group->free_space += info->bytes;
-	block_group->free_extents++;
+	ctl->free_space += info->bytes;
+	ctl->free_extents++;
 	return ret;
 }
 
-static void recalculate_thresholds(struct btrfs_block_group_cache *block_group)
+static void recalculate_thresholds(struct btrfs_free_space_ctl *ctl)
 {
+	struct btrfs_block_group_cache *block_group = ctl->private;
 	u64 max_bytes;
 	u64 bitmap_bytes;
 	u64 extent_bytes;
 	u64 size = block_group->key.offset;
+	u64 bytes_per_bg = BITS_PER_BITMAP * block_group->sectorsize;
+	int max_bitmaps = div64_u64(size + bytes_per_bg - 1, bytes_per_bg);
+
+	BUG_ON(ctl->total_bitmaps > max_bitmaps);
 
 	/*
 	 * The goal is to keep the total amount of memory used per 1gb of space
@@ -1112,10 +1170,10 @@ static void recalculate_thresholds(struct btrfs_block_group_cache *block_group)
 	 * sure we don't go over our overall goal of MAX_CACHE_BYTES_PER_GIG as
 	 * we add more bitmaps.
 	 */
-	bitmap_bytes = (block_group->total_bitmaps + 1) * PAGE_CACHE_SIZE;
+	bitmap_bytes = (ctl->total_bitmaps + 1) * PAGE_CACHE_SIZE;
 
 	if (bitmap_bytes >= max_bytes) {
-		block_group->extents_thresh = 0;
+		ctl->extents_thresh = 0;
 		return;
 	}
 
@@ -1126,47 +1184,43 @@ static void recalculate_thresholds(struct btrfs_block_group_cache *block_group)
 	extent_bytes = max_bytes - bitmap_bytes;
 	extent_bytes = min_t(u64, extent_bytes, div64_u64(max_bytes, 2));
 
-	block_group->extents_thresh =
+	ctl->extents_thresh =
 		div64_u64(extent_bytes, (sizeof(struct btrfs_free_space)));
 }
 
-static void bitmap_clear_bits(struct btrfs_block_group_cache *block_group,
+static void bitmap_clear_bits(struct btrfs_free_space_ctl *ctl,
 			      struct btrfs_free_space *info, u64 offset,
 			      u64 bytes)
 {
-	unsigned long start, end;
-	unsigned long i;
+	unsigned long start, count;
 
-	start = offset_to_bit(info->offset, block_group->sectorsize, offset);
-	end = start + bytes_to_bits(bytes, block_group->sectorsize);
-	BUG_ON(end > BITS_PER_BITMAP);
+	start = offset_to_bit(info->offset, ctl->unit, offset);
+	count = bytes_to_bits(bytes, ctl->unit);
+	BUG_ON(start + count > BITS_PER_BITMAP);
 
-	for (i = start; i < end; i++)
-		clear_bit(i, info->bitmap);
+	bitmap_clear(info->bitmap, start, count);
 
 	info->bytes -= bytes;
-	block_group->free_space -= bytes;
+	ctl->free_space -= bytes;
 }
 
-static void bitmap_set_bits(struct btrfs_block_group_cache *block_group,
+static void bitmap_set_bits(struct btrfs_free_space_ctl *ctl,
 			    struct btrfs_free_space *info, u64 offset,
 			    u64 bytes)
 {
-	unsigned long start, end;
-	unsigned long i;
+	unsigned long start, count;
 
-	start = offset_to_bit(info->offset, block_group->sectorsize, offset);
-	end = start + bytes_to_bits(bytes, block_group->sectorsize);
-	BUG_ON(end > BITS_PER_BITMAP);
+	start = offset_to_bit(info->offset, ctl->unit, offset);
+	count = bytes_to_bits(bytes, ctl->unit);
+	BUG_ON(start + count > BITS_PER_BITMAP);
 
-	for (i = start; i < end; i++)
-		set_bit(i, info->bitmap);
+	bitmap_set(info->bitmap, start, count);
 
 	info->bytes += bytes;
-	block_group->free_space += bytes;
+	ctl->free_space += bytes;
 }
 
-static int search_bitmap(struct btrfs_block_group_cache *block_group,
+static int search_bitmap(struct btrfs_free_space_ctl *ctl,
 			 struct btrfs_free_space *bitmap_info, u64 *offset,
 			 u64 *bytes)
 {
@@ -1174,9 +1228,9 @@ static int search_bitmap(struct btrfs_block_group_cache *block_group,
 	unsigned long bits, i;
 	unsigned long next_zero;
 
-	i = offset_to_bit(bitmap_info->offset, block_group->sectorsize,
+	i = offset_to_bit(bitmap_info->offset, ctl->unit,
 			  max_t(u64, *offset, bitmap_info->offset));
-	bits = bytes_to_bits(*bytes, block_group->sectorsize);
+	bits = bytes_to_bits(*bytes, ctl->unit);
 
 	for (i = find_next_bit(bitmap_info->bitmap, BITS_PER_BITMAP, i);
 	     i < BITS_PER_BITMAP;
@@ -1191,29 +1245,25 @@ static int search_bitmap(struct btrfs_block_group_cache *block_group,
 	}
 
 	if (found_bits) {
-		*offset = (u64)(i * block_group->sectorsize) +
-			bitmap_info->offset;
-		*bytes = (u64)(found_bits) * block_group->sectorsize;
+		*offset = (u64)(i * ctl->unit) + bitmap_info->offset;
+		*bytes = (u64)(found_bits) * ctl->unit;
 		return 0;
 	}
 
 	return -1;
 }
 
-static struct btrfs_free_space *find_free_space(struct btrfs_block_group_cache
-						*block_group, u64 *offset,
-						u64 *bytes, int debug)
+static struct btrfs_free_space *
+find_free_space(struct btrfs_free_space_ctl *ctl, u64 *offset, u64 *bytes)
 {
 	struct btrfs_free_space *entry;
 	struct rb_node *node;
 	int ret;
 
-	if (!block_group->free_space_offset.rb_node)
+	if (!ctl->free_space_offset.rb_node)
 		return NULL;
 
-	entry = tree_search_offset(block_group,
-				   offset_to_bitmap(block_group, *offset),
-				   0, 1);
+	entry = tree_search_offset(ctl, offset_to_bitmap(ctl, *offset), 0, 1);
 	if (!entry)
 		return NULL;
 
@@ -1223,7 +1273,7 @@ static struct btrfs_free_space *find_free_space(struct btrfs_block_group_cache
 			continue;
 
 		if (entry->bitmap) {
-			ret = search_bitmap(block_group, entry, offset, bytes);
+			ret = search_bitmap(ctl, entry, offset, bytes);
 			if (!ret)
 				return entry;
 			continue;
@@ -1237,33 +1287,28 @@ static struct btrfs_free_space *find_free_space(struct btrfs_block_group_cache
 	return NULL;
 }
 
-static void add_new_bitmap(struct btrfs_block_group_cache *block_group,
+static void add_new_bitmap(struct btrfs_free_space_ctl *ctl,
 			   struct btrfs_free_space *info, u64 offset)
 {
-	u64 bytes_per_bg = BITS_PER_BITMAP * block_group->sectorsize;
-	int max_bitmaps = (int)div64_u64(block_group->key.offset +
-					 bytes_per_bg - 1, bytes_per_bg);
-	BUG_ON(block_group->total_bitmaps >= max_bitmaps);
-
-	info->offset = offset_to_bitmap(block_group, offset);
+	info->offset = offset_to_bitmap(ctl, offset);
 	info->bytes = 0;
-	link_free_space(block_group, info);
-	block_group->total_bitmaps++;
+	link_free_space(ctl, info);
+	ctl->total_bitmaps++;
 
-	recalculate_thresholds(block_group);
+	ctl->op->recalc_thresholds(ctl);
 }
 
-static void free_bitmap(struct btrfs_block_group_cache *block_group,
+static void free_bitmap(struct btrfs_free_space_ctl *ctl,
 			struct btrfs_free_space *bitmap_info)
 {
-	unlink_free_space(block_group, bitmap_info);
+	unlink_free_space(ctl, bitmap_info);
 	kfree(bitmap_info->bitmap);
 	kmem_cache_free(btrfs_free_space_cachep, bitmap_info);
-	block_group->total_bitmaps--;
-	recalculate_thresholds(block_group);
+	ctl->total_bitmaps--;
+	ctl->op->recalc_thresholds(ctl);
 }
 
-static noinline int remove_from_bitmap(struct btrfs_block_group_cache *block_group,
+static noinline int remove_from_bitmap(struct btrfs_free_space_ctl *ctl,
 			      struct btrfs_free_space *bitmap_info,
 			      u64 *offset, u64 *bytes)
 {
@@ -1272,8 +1317,7 @@ static noinline int remove_from_bitmap(struct btrfs_block_group_cache *block_gro
 	int ret;
 
 again:
-	end = bitmap_info->offset +
-		(u64)(BITS_PER_BITMAP * block_group->sectorsize) - 1;
+	end = bitmap_info->offset + (u64)(BITS_PER_BITMAP * ctl->unit) - 1;
 
 	/*
 	 * XXX - this can go away after a few releases.
@@ -1288,24 +1332,22 @@ again:
 	search_start = *offset;
 	search_bytes = *bytes;
 	search_bytes = min(search_bytes, end - search_start + 1);
-	ret = search_bitmap(block_group, bitmap_info, &search_start,
-			    &search_bytes);
+	ret = search_bitmap(ctl, bitmap_info, &search_start, &search_bytes);
 	BUG_ON(ret < 0 || search_start != *offset);
 
 	if (*offset > bitmap_info->offset && *offset + *bytes > end) {
-		bitmap_clear_bits(block_group, bitmap_info, *offset,
-				  end - *offset + 1);
+		bitmap_clear_bits(ctl, bitmap_info, *offset, end - *offset + 1);
 		*bytes -= end - *offset + 1;
 		*offset = end + 1;
 	} else if (*offset >= bitmap_info->offset && *offset + *bytes <= end) {
-		bitmap_clear_bits(block_group, bitmap_info, *offset, *bytes);
+		bitmap_clear_bits(ctl, bitmap_info, *offset, *bytes);
 		*bytes = 0;
 	}
 
 	if (*bytes) {
 		struct rb_node *next = rb_next(&bitmap_info->offset_index);
 		if (!bitmap_info->bytes)
-			free_bitmap(block_group, bitmap_info);
+			free_bitmap(ctl, bitmap_info);
 
 		/*
 		 * no entry after this bitmap, but we still have bytes to
@@ -1332,31 +1374,28 @@ again:
 		 */
 		search_start = *offset;
 		search_bytes = *bytes;
-		ret = search_bitmap(block_group, bitmap_info, &search_start,
+		ret = search_bitmap(ctl, bitmap_info, &search_start,
 				    &search_bytes);
 		if (ret < 0 || search_start != *offset)
 			return -EAGAIN;
 
 		goto again;
 	} else if (!bitmap_info->bytes)
-		free_bitmap(block_group, bitmap_info);
+		free_bitmap(ctl, bitmap_info);
 
 	return 0;
 }
 
-static int insert_into_bitmap(struct btrfs_block_group_cache *block_group,
-			      struct btrfs_free_space *info)
+static bool use_bitmap(struct btrfs_free_space_ctl *ctl,
+		      struct btrfs_free_space *info)
 {
-	struct btrfs_free_space *bitmap_info;
-	int added = 0;
-	u64 bytes, offset, end;
-	int ret;
+	struct btrfs_block_group_cache *block_group = ctl->private;
 
 	/*
 	 * If we are below the extents threshold then we can add this as an
 	 * extent, and don't have to deal with the bitmap
 	 */
-	if (block_group->free_extents < block_group->extents_thresh) {
+	if (ctl->free_extents < ctl->extents_thresh) {
 		/*
 		 * If this block group has some small extents we don't want to
 		 * use up all of our free slots in the cache with them, we want
@@ -1365,11 +1404,10 @@ static int insert_into_bitmap(struct btrfs_block_group_cache *block_group,
 		 * the overhead of a bitmap if we don't have to.
 		 */
 		if (info->bytes <= block_group->sectorsize * 4) {
-			if (block_group->free_extents * 2 <=
-			    block_group->extents_thresh)
-				return 0;
+			if (ctl->free_extents * 2 <= ctl->extents_thresh)
+				return false;
 		} else {
-			return 0;
+			return false;
 		}
 	}
 
@@ -1379,31 +1417,42 @@ static int insert_into_bitmap(struct btrfs_block_group_cache *block_group,
 	 */
 	if (BITS_PER_BITMAP * block_group->sectorsize >
 	    block_group->key.offset)
-		return 0;
+		return false;
+
+	return true;
+}
+
+static int insert_into_bitmap(struct btrfs_free_space_ctl *ctl,
+			      struct btrfs_free_space *info)
+{
+	struct btrfs_free_space *bitmap_info;
+	int added = 0;
+	u64 bytes, offset, end;
+	int ret;
 
 	bytes = info->bytes;
 	offset = info->offset;
 
+	if (!ctl->op->use_bitmap(ctl, info))
+		return 0;
+
 again:
-	bitmap_info = tree_search_offset(block_group,
-					 offset_to_bitmap(block_group, offset),
+	bitmap_info = tree_search_offset(ctl, offset_to_bitmap(ctl, offset),
 					 1, 0);
 	if (!bitmap_info) {
 		BUG_ON(added);
 		goto new_bitmap;
 	}
 
-	end = bitmap_info->offset +
-		(u64)(BITS_PER_BITMAP * block_group->sectorsize);
+	end = bitmap_info->offset + (u64)(BITS_PER_BITMAP * ctl->unit);
 
 	if (offset >= bitmap_info->offset && offset + bytes > end) {
-		bitmap_set_bits(block_group, bitmap_info, offset,
-				end - offset);
+		bitmap_set_bits(ctl, bitmap_info, offset, end - offset);
 		bytes -= end - offset;
 		offset = end;
 		added = 0;
 	} else if (offset >= bitmap_info->offset && offset + bytes <= end) {
-		bitmap_set_bits(block_group, bitmap_info, offset, bytes);
+		bitmap_set_bits(ctl, bitmap_info, offset, bytes);
 		bytes = 0;
 	} else {
 		BUG();
@@ -1417,19 +1466,19 @@ again:
 
 new_bitmap:
 	if (info && info->bitmap) {
-		add_new_bitmap(block_group, info, offset);
+		add_new_bitmap(ctl, info, offset);
 		added = 1;
 		info = NULL;
 		goto again;
 	} else {
-		spin_unlock(&block_group->tree_lock);
+		spin_unlock(&ctl->tree_lock);
 
 		/* no pre-allocated info, allocate a new one */
 		if (!info) {
 			info = kmem_cache_zalloc(btrfs_free_space_cachep,
 						 GFP_NOFS);
 			if (!info) {
-				spin_lock(&block_group->tree_lock);
+				spin_lock(&ctl->tree_lock);
 				ret = -ENOMEM;
 				goto out;
 			}
@@ -1437,7 +1486,7 @@ new_bitmap:
 
 		/* allocate the bitmap */
 		info->bitmap = kzalloc(PAGE_CACHE_SIZE, GFP_NOFS);
-		spin_lock(&block_group->tree_lock);
+		spin_lock(&ctl->tree_lock);
 		if (!info->bitmap) {
 			ret = -ENOMEM;
 			goto out;
@@ -1455,7 +1504,7 @@ out:
 	return ret;
 }
 
-bool try_merge_free_space(struct btrfs_block_group_cache *block_group,
+bool try_merge_free_space(struct btrfs_free_space_ctl *ctl,
 			  struct btrfs_free_space *info, bool update_stat)
 {
 	struct btrfs_free_space *left_info;
@@ -1469,18 +1518,18 @@ bool try_merge_free_space(struct btrfs_block_group_cache *block_group,
 	 * are adding, if there is remove that struct and add a new one to
 	 * cover the entire range
 	 */
-	right_info = tree_search_offset(block_group, offset + bytes, 0, 0);
+	right_info = tree_search_offset(ctl, offset + bytes, 0, 0);
 	if (right_info && rb_prev(&right_info->offset_index))
 		left_info = rb_entry(rb_prev(&right_info->offset_index),
 				     struct btrfs_free_space, offset_index);
 	else
-		left_info = tree_search_offset(block_group, offset - 1, 0, 0);
+		left_info = tree_search_offset(ctl, offset - 1, 0, 0);
 
 	if (right_info && !right_info->bitmap) {
 		if (update_stat)
-			unlink_free_space(block_group, right_info);
+			unlink_free_space(ctl, right_info);
 		else
-			__unlink_free_space(block_group, right_info);
+			__unlink_free_space(ctl, right_info);
 		info->bytes += right_info->bytes;
 		kmem_cache_free(btrfs_free_space_cachep, right_info);
 		merged = true;
@@ -1489,9 +1538,9 @@ bool try_merge_free_space(struct btrfs_block_group_cache *block_group,
 	if (left_info && !left_info->bitmap &&
 	    left_info->offset + left_info->bytes == offset) {
 		if (update_stat)
-			unlink_free_space(block_group, left_info);
+			unlink_free_space(ctl, left_info);
 		else
-			__unlink_free_space(block_group, left_info);
+			__unlink_free_space(ctl, left_info);
 		info->offset = left_info->offset;
 		info->bytes += left_info->bytes;
 		kmem_cache_free(btrfs_free_space_cachep, left_info);
@@ -1501,8 +1550,8 @@ bool try_merge_free_space(struct btrfs_block_group_cache *block_group,
 	return merged;
 }
 
-int btrfs_add_free_space(struct btrfs_block_group_cache *block_group,
-			 u64 offset, u64 bytes)
+int __btrfs_add_free_space(struct btrfs_free_space_ctl *ctl,
+			   u64 offset, u64 bytes)
 {
 	struct btrfs_free_space *info;
 	int ret = 0;
@@ -1514,9 +1563,9 @@ int btrfs_add_free_space(struct btrfs_block_group_cache *block_group,
 	info->offset = offset;
 	info->bytes = bytes;
 
-	spin_lock(&block_group->tree_lock);
+	spin_lock(&ctl->tree_lock);
 
-	if (try_merge_free_space(block_group, info, true))
+	if (try_merge_free_space(ctl, info, true))
 		goto link;
 
 	/*
@@ -1524,7 +1573,7 @@ int btrfs_add_free_space(struct btrfs_block_group_cache *block_group,
 	 * extent then we know we're going to have to allocate a new extent, so
 	 * before we do that see if we need to drop this into a bitmap
 	 */
-	ret = insert_into_bitmap(block_group, info);
+	ret = insert_into_bitmap(ctl, info);
 	if (ret < 0) {
 		goto out;
 	} else if (ret) {
@@ -1532,11 +1581,11 @@ int btrfs_add_free_space(struct btrfs_block_group_cache *block_group,
 		goto out;
 	}
 link:
-	ret = link_free_space(block_group, info);
+	ret = link_free_space(ctl, info);
 	if (ret)
 		kmem_cache_free(btrfs_free_space_cachep, info);
 out:
-	spin_unlock(&block_group->tree_lock);
+	spin_unlock(&ctl->tree_lock);
 
 	if (ret) {
 		printk(KERN_CRIT "btrfs: unable to add free space :%d\n", ret);
@@ -1549,21 +1598,21 @@ out:
 int btrfs_remove_free_space(struct btrfs_block_group_cache *block_group,
 			    u64 offset, u64 bytes)
 {
+	struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl;
 	struct btrfs_free_space *info;
 	struct btrfs_free_space *next_info = NULL;
 	int ret = 0;
 
-	spin_lock(&block_group->tree_lock);
+	spin_lock(&ctl->tree_lock);
 
 again:
-	info = tree_search_offset(block_group, offset, 0, 0);
+	info = tree_search_offset(ctl, offset, 0, 0);
 	if (!info) {
 		/*
 		 * oops didn't find an extent that matched the space we wanted
 		 * to remove, look for a bitmap instead
 		 */
-		info = tree_search_offset(block_group,
-					  offset_to_bitmap(block_group, offset),
+		info = tree_search_offset(ctl, offset_to_bitmap(ctl, offset),
 					  1, 0);
 		if (!info) {
 			WARN_ON(1);
@@ -1578,8 +1627,8 @@ again:
 					     offset_index);
 
 		if (next_info->bitmap)
-			end = next_info->offset + BITS_PER_BITMAP *
-				block_group->sectorsize - 1;
+			end = next_info->offset +
+			      BITS_PER_BITMAP * ctl->unit - 1;
 		else
 			end = next_info->offset + next_info->bytes;
 
@@ -1599,20 +1648,20 @@ again:
 	}
 
 	if (info->bytes == bytes) {
-		unlink_free_space(block_group, info);
+		unlink_free_space(ctl, info);
 		if (info->bitmap) {
 			kfree(info->bitmap);
-			block_group->total_bitmaps--;
+			ctl->total_bitmaps--;
 		}
 		kmem_cache_free(btrfs_free_space_cachep, info);
 		goto out_lock;
 	}
 
 	if (!info->bitmap && info->offset == offset) {
-		unlink_free_space(block_group, info);
+		unlink_free_space(ctl, info);
 		info->offset += bytes;
 		info->bytes -= bytes;
-		link_free_space(block_group, info);
+		link_free_space(ctl, info);
 		goto out_lock;
 	}
 
@@ -1626,13 +1675,13 @@ again:
 		 * first unlink the old info and then
 		 * insert it again after the hole we're creating
 		 */
-		unlink_free_space(block_group, info);
+		unlink_free_space(ctl, info);
 		if (offset + bytes < info->offset + info->bytes) {
 			u64 old_end = info->offset + info->bytes;
 
 			info->offset = offset + bytes;
 			info->bytes = old_end - info->offset;
-			ret = link_free_space(block_group, info);
+			ret = link_free_space(ctl, info);
 			WARN_ON(ret);
 			if (ret)
 				goto out_lock;
@@ -1642,7 +1691,7 @@ again:
 			 */
 			kmem_cache_free(btrfs_free_space_cachep, info);
 		}
-		spin_unlock(&block_group->tree_lock);
+		spin_unlock(&ctl->tree_lock);
 
 		/* step two, insert a new info struct to cover
 		 * anything before the hole
@@ -1653,12 +1702,12 @@ again:
 		goto out;
 	}
 
-	ret = remove_from_bitmap(block_group, info, &offset, &bytes);
+	ret = remove_from_bitmap(ctl, info, &offset, &bytes);
 	if (ret == -EAGAIN)
 		goto again;
 	BUG_ON(ret);
 out_lock:
-	spin_unlock(&block_group->tree_lock);
+	spin_unlock(&ctl->tree_lock);
 out:
 	return ret;
 }
@@ -1666,11 +1715,12 @@ out:
 void btrfs_dump_free_space(struct btrfs_block_group_cache *block_group,
 			   u64 bytes)
 {
+	struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl;
 	struct btrfs_free_space *info;
 	struct rb_node *n;
 	int count = 0;
 
-	for (n = rb_first(&block_group->free_space_offset); n; n = rb_next(n)) {
+	for (n = rb_first(&ctl->free_space_offset); n; n = rb_next(n)) {
 		info = rb_entry(n, struct btrfs_free_space, offset_index);
 		if (info->bytes >= bytes)
 			count++;
@@ -1685,19 +1735,28 @@ void btrfs_dump_free_space(struct btrfs_block_group_cache *block_group,
 	       "\n", count);
 }
 
-u64 btrfs_block_group_free_space(struct btrfs_block_group_cache *block_group)
+static struct btrfs_free_space_op free_space_op = {
+	.recalc_thresholds	= recalculate_thresholds,
+	.use_bitmap		= use_bitmap,
+};
+
+void btrfs_init_free_space_ctl(struct btrfs_block_group_cache *block_group)
 {
-	struct btrfs_free_space *info;
-	struct rb_node *n;
-	u64 ret = 0;
+	struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl;
 
-	for (n = rb_first(&block_group->free_space_offset); n;
-	     n = rb_next(n)) {
-		info = rb_entry(n, struct btrfs_free_space, offset_index);
-		ret += info->bytes;
-	}
+	spin_lock_init(&ctl->tree_lock);
+	ctl->unit = block_group->sectorsize;
+	ctl->start = block_group->key.objectid;
+	ctl->private = block_group;
+	ctl->op = &free_space_op;
 
-	return ret;
+	/*
+	 * we only want to have 32k of ram per block group for keeping
+	 * track of free space, and if we pass 1/2 of that we want to
+	 * start converting things over to using bitmaps
+	 */
+	ctl->extents_thresh = ((1024 * 32) / 2) /
+				sizeof(struct btrfs_free_space);
 }
 
 /*
@@ -1711,6 +1770,7 @@ __btrfs_return_cluster_to_free_space(
 			     struct btrfs_block_group_cache *block_group,
 			     struct btrfs_free_cluster *cluster)
 {
+	struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl;
 	struct btrfs_free_space *entry;
 	struct rb_node *node;
 
@@ -1732,8 +1792,8 @@ __btrfs_return_cluster_to_free_space(
 
 		bitmap = (entry->bitmap != NULL);
 		if (!bitmap)
-			try_merge_free_space(block_group, entry, false);
-		tree_insert_offset(&block_group->free_space_offset,
+			try_merge_free_space(ctl, entry, false);
+		tree_insert_offset(&ctl->free_space_offset,
 				   entry->offset, &entry->offset_index, bitmap);
 	}
 	cluster->root = RB_ROOT;
@@ -1744,14 +1804,38 @@ out:
 	return 0;
 }
 
-void btrfs_remove_free_space_cache(struct btrfs_block_group_cache *block_group)
+void __btrfs_remove_free_space_cache_locked(struct btrfs_free_space_ctl *ctl)
 {
 	struct btrfs_free_space *info;
 	struct rb_node *node;
+
+	while ((node = rb_last(&ctl->free_space_offset)) != NULL) {
+		info = rb_entry(node, struct btrfs_free_space, offset_index);
+		unlink_free_space(ctl, info);
+		kfree(info->bitmap);
+		kmem_cache_free(btrfs_free_space_cachep, info);
+		if (need_resched()) {
+			spin_unlock(&ctl->tree_lock);
+			cond_resched();
+			spin_lock(&ctl->tree_lock);
+		}
+	}
+}
+
+void __btrfs_remove_free_space_cache(struct btrfs_free_space_ctl *ctl)
+{
+	spin_lock(&ctl->tree_lock);
+	__btrfs_remove_free_space_cache_locked(ctl);
+	spin_unlock(&ctl->tree_lock);
+}
+
+void btrfs_remove_free_space_cache(struct btrfs_block_group_cache *block_group)
+{
+	struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl;
 	struct btrfs_free_cluster *cluster;
 	struct list_head *head;
 
-	spin_lock(&block_group->tree_lock);
+	spin_lock(&ctl->tree_lock);
 	while ((head = block_group->cluster_list.next) !=
 	       &block_group->cluster_list) {
 		cluster = list_entry(head, struct btrfs_free_cluster,
@@ -1760,60 +1844,46 @@ void btrfs_remove_free_space_cache(struct btrfs_block_group_cache *block_group)
 		WARN_ON(cluster->block_group != block_group);
 		__btrfs_return_cluster_to_free_space(block_group, cluster);
 		if (need_resched()) {
-			spin_unlock(&block_group->tree_lock);
+			spin_unlock(&ctl->tree_lock);
 			cond_resched();
-			spin_lock(&block_group->tree_lock);
+			spin_lock(&ctl->tree_lock);
 		}
 	}
+	__btrfs_remove_free_space_cache_locked(ctl);
+	spin_unlock(&ctl->tree_lock);
 
-	while ((node = rb_last(&block_group->free_space_offset)) != NULL) {
-		info = rb_entry(node, struct btrfs_free_space, offset_index);
-		if (!info->bitmap) {
-			unlink_free_space(block_group, info);
-			kmem_cache_free(btrfs_free_space_cachep, info);
-		} else {
-			free_bitmap(block_group, info);
-		}
-
-		if (need_resched()) {
-			spin_unlock(&block_group->tree_lock);
-			cond_resched();
-			spin_lock(&block_group->tree_lock);
-		}
-	}
-
-	spin_unlock(&block_group->tree_lock);
 }
 
 u64 btrfs_find_space_for_alloc(struct btrfs_block_group_cache *block_group,
 			       u64 offset, u64 bytes, u64 empty_size)
 {
+	struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl;
 	struct btrfs_free_space *entry = NULL;
 	u64 bytes_search = bytes + empty_size;
 	u64 ret = 0;
 
-	spin_lock(&block_group->tree_lock);
-	entry = find_free_space(block_group, &offset, &bytes_search, 0);
+	spin_lock(&ctl->tree_lock);
+	entry = find_free_space(ctl, &offset, &bytes_search);
 	if (!entry)
 		goto out;
 
 	ret = offset;
 	if (entry->bitmap) {
-		bitmap_clear_bits(block_group, entry, offset, bytes);
+		bitmap_clear_bits(ctl, entry, offset, bytes);
 		if (!entry->bytes)
-			free_bitmap(block_group, entry);
+			free_bitmap(ctl, entry);
 	} else {
-		unlink_free_space(block_group, entry);
+		unlink_free_space(ctl, entry);
 		entry->offset += bytes;
 		entry->bytes -= bytes;
 		if (!entry->bytes)
 			kmem_cache_free(btrfs_free_space_cachep, entry);
 		else
-			link_free_space(block_group, entry);
+			link_free_space(ctl, entry);
 	}
 
 out:
-	spin_unlock(&block_group->tree_lock);
+	spin_unlock(&ctl->tree_lock);
 
 	return ret;
 }
@@ -1830,6 +1900,7 @@ int btrfs_return_cluster_to_free_space(
 			       struct btrfs_block_group_cache *block_group,
 			       struct btrfs_free_cluster *cluster)
 {
+	struct btrfs_free_space_ctl *ctl;
 	int ret;
 
 	/* first, get a safe pointer to the block group */
@@ -1848,10 +1919,12 @@ int btrfs_return_cluster_to_free_space(
 	atomic_inc(&block_group->count);
 	spin_unlock(&cluster->lock);
 
+	ctl = block_group->free_space_ctl;
+
 	/* now return any extents the cluster had on it */
-	spin_lock(&block_group->tree_lock);
+	spin_lock(&ctl->tree_lock);
 	ret = __btrfs_return_cluster_to_free_space(block_group, cluster);
-	spin_unlock(&block_group->tree_lock);
+	spin_unlock(&ctl->tree_lock);
 
 	/* finally drop our ref */
 	btrfs_put_block_group(block_group);
@@ -1863,6 +1936,7 @@ static u64 btrfs_alloc_from_bitmap(struct btrfs_block_group_cache *block_group,
 				   struct btrfs_free_space *entry,
 				   u64 bytes, u64 min_start)
 {
+	struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl;
 	int err;
 	u64 search_start = cluster->window_start;
 	u64 search_bytes = bytes;
@@ -1871,13 +1945,12 @@ static u64 btrfs_alloc_from_bitmap(struct btrfs_block_group_cache *block_group,
 	search_start = min_start;
 	search_bytes = bytes;
 
-	err = search_bitmap(block_group, entry, &search_start,
-			    &search_bytes);
+	err = search_bitmap(ctl, entry, &search_start, &search_bytes);
 	if (err)
 		return 0;
 
 	ret = search_start;
-	bitmap_clear_bits(block_group, entry, ret, bytes);
+	bitmap_clear_bits(ctl, entry, ret, bytes);
 
 	return ret;
 }
@@ -1891,6 +1964,7 @@ u64 btrfs_alloc_from_cluster(struct btrfs_block_group_cache *block_group,
 			     struct btrfs_free_cluster *cluster, u64 bytes,
 			     u64 min_start)
 {
+	struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl;
 	struct btrfs_free_space *entry = NULL;
 	struct rb_node *node;
 	u64 ret = 0;
@@ -1951,20 +2025,20 @@ out:
 	if (!ret)
 		return 0;
 
-	spin_lock(&block_group->tree_lock);
+	spin_lock(&ctl->tree_lock);
 
-	block_group->free_space -= bytes;
+	ctl->free_space -= bytes;
 	if (entry->bytes == 0) {
-		block_group->free_extents--;
+		ctl->free_extents--;
 		if (entry->bitmap) {
 			kfree(entry->bitmap);
-			block_group->total_bitmaps--;
-			recalculate_thresholds(block_group);
+			ctl->total_bitmaps--;
+			ctl->op->recalc_thresholds(ctl);
 		}
 		kmem_cache_free(btrfs_free_space_cachep, entry);
 	}
 
-	spin_unlock(&block_group->tree_lock);
+	spin_unlock(&ctl->tree_lock);
 
 	return ret;
 }
@@ -1974,6 +2048,7 @@ static int btrfs_bitmap_cluster(struct btrfs_block_group_cache *block_group,
 				struct btrfs_free_cluster *cluster,
 				u64 offset, u64 bytes, u64 min_bytes)
 {
+	struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl;
 	unsigned long next_zero;
 	unsigned long i;
 	unsigned long search_bits;
@@ -2028,7 +2103,7 @@ again:
 
 	cluster->window_start = start * block_group->sectorsize +
 		entry->offset;
-	rb_erase(&entry->offset_index, &block_group->free_space_offset);
+	rb_erase(&entry->offset_index, &ctl->free_space_offset);
 	ret = tree_insert_offset(&cluster->root, entry->offset,
 				 &entry->offset_index, 1);
 	BUG_ON(ret);
@@ -2043,6 +2118,7 @@ static int setup_cluster_no_bitmap(struct btrfs_block_group_cache *block_group,
 				   struct btrfs_free_cluster *cluster,
 				   u64 offset, u64 bytes, u64 min_bytes)
 {
+	struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl;
 	struct btrfs_free_space *first = NULL;
 	struct btrfs_free_space *entry = NULL;
 	struct btrfs_free_space *prev = NULL;
@@ -2053,7 +2129,7 @@ static int setup_cluster_no_bitmap(struct btrfs_block_group_cache *block_group,
 	u64 max_extent;
 	u64 max_gap = 128 * 1024;
 
-	entry = tree_search_offset(block_group, offset, 0, 1);
+	entry = tree_search_offset(ctl, offset, 0, 1);
 	if (!entry)
 		return -ENOSPC;
 
@@ -2119,7 +2195,7 @@ static int setup_cluster_no_bitmap(struct btrfs_block_group_cache *block_group,
 		if (entry->bitmap)
 			continue;
 
-		rb_erase(&entry->offset_index, &block_group->free_space_offset);
+		rb_erase(&entry->offset_index, &ctl->free_space_offset);
 		ret = tree_insert_offset(&cluster->root, entry->offset,
 					 &entry->offset_index, 0);
 		BUG_ON(ret);
@@ -2138,16 +2214,15 @@ static int setup_cluster_bitmap(struct btrfs_block_group_cache *block_group,
 				struct btrfs_free_cluster *cluster,
 				u64 offset, u64 bytes, u64 min_bytes)
 {
+	struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl;
 	struct btrfs_free_space *entry;
 	struct rb_node *node;
 	int ret = -ENOSPC;
 
-	if (block_group->total_bitmaps == 0)
+	if (ctl->total_bitmaps == 0)
 		return -ENOSPC;
 
-	entry = tree_search_offset(block_group,
-				   offset_to_bitmap(block_group, offset),
-				   0, 1);
+	entry = tree_search_offset(ctl, offset_to_bitmap(ctl, offset), 0, 1);
 	if (!entry)
 		return -ENOSPC;
 
@@ -2180,6 +2255,7 @@ int btrfs_find_space_cluster(struct btrfs_trans_handle *trans,
 			     struct btrfs_free_cluster *cluster,
 			     u64 offset, u64 bytes, u64 empty_size)
 {
+	struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl;
 	u64 min_bytes;
 	int ret;
 
@@ -2199,14 +2275,14 @@ int btrfs_find_space_cluster(struct btrfs_trans_handle *trans,
 	} else
 		min_bytes = max(bytes, (bytes + empty_size) >> 2);
 
-	spin_lock(&block_group->tree_lock);
+	spin_lock(&ctl->tree_lock);
 
 	/*
 	 * If we know we don't have enough space to make a cluster don't even
 	 * bother doing all the work to try and find one.
 	 */
-	if (block_group->free_space < min_bytes) {
-		spin_unlock(&block_group->tree_lock);
+	if (ctl->free_space < min_bytes) {
+		spin_unlock(&ctl->tree_lock);
 		return -ENOSPC;
 	}
 
@@ -2232,7 +2308,7 @@ int btrfs_find_space_cluster(struct btrfs_trans_handle *trans,
 	}
 out:
 	spin_unlock(&cluster->lock);
-	spin_unlock(&block_group->tree_lock);
+	spin_unlock(&ctl->tree_lock);
 
 	return ret;
 }
@@ -2253,6 +2329,7 @@ void btrfs_init_free_cluster(struct btrfs_free_cluster *cluster)
 int btrfs_trim_block_group(struct btrfs_block_group_cache *block_group,
 			   u64 *trimmed, u64 start, u64 end, u64 minlen)
 {
+	struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl;
 	struct btrfs_free_space *entry = NULL;
 	struct btrfs_fs_info *fs_info = block_group->fs_info;
 	u64 bytes = 0;
@@ -2262,52 +2339,50 @@ int btrfs_trim_block_group(struct btrfs_block_group_cache *block_group,
 	*trimmed = 0;
 
 	while (start < end) {
-		spin_lock(&block_group->tree_lock);
+		spin_lock(&ctl->tree_lock);
 
-		if (block_group->free_space < minlen) {
-			spin_unlock(&block_group->tree_lock);
+		if (ctl->free_space < minlen) {
+			spin_unlock(&ctl->tree_lock);
 			break;
 		}
 
-		entry = tree_search_offset(block_group, start, 0, 1);
+		entry = tree_search_offset(ctl, start, 0, 1);
 		if (!entry)
-			entry = tree_search_offset(block_group,
-						   offset_to_bitmap(block_group,
-								    start),
+			entry = tree_search_offset(ctl,
+						   offset_to_bitmap(ctl, start),
 						   1, 1);
 
 		if (!entry || entry->offset >= end) {
-			spin_unlock(&block_group->tree_lock);
+			spin_unlock(&ctl->tree_lock);
 			break;
 		}
 
 		if (entry->bitmap) {
-			ret = search_bitmap(block_group, entry, &start, &bytes);
+			ret = search_bitmap(ctl, entry, &start, &bytes);
 			if (!ret) {
 				if (start >= end) {
-					spin_unlock(&block_group->tree_lock);
+					spin_unlock(&ctl->tree_lock);
 					break;
 				}
 				bytes = min(bytes, end - start);
-				bitmap_clear_bits(block_group, entry,
-						  start, bytes);
+				bitmap_clear_bits(ctl, entry, start, bytes);
 				if (entry->bytes == 0)
-					free_bitmap(block_group, entry);
+					free_bitmap(ctl, entry);
 			} else {
 				start = entry->offset + BITS_PER_BITMAP *
 					block_group->sectorsize;
-				spin_unlock(&block_group->tree_lock);
+				spin_unlock(&ctl->tree_lock);
 				ret = 0;
 				continue;
 			}
 		} else {
 			start = entry->offset;
 			bytes = min(entry->bytes, end - start);
-			unlink_free_space(block_group, entry);
+			unlink_free_space(ctl, entry);
 			kmem_cache_free(btrfs_free_space_cachep, entry);
 		}
 
-		spin_unlock(&block_group->tree_lock);
+		spin_unlock(&ctl->tree_lock);
 
 		if (bytes >= minlen) {
 			int update_ret;
@@ -2319,8 +2394,7 @@ int btrfs_trim_block_group(struct btrfs_block_group_cache *block_group,
 							 bytes,
 							 &actually_trimmed);
 
-			btrfs_add_free_space(block_group,
-					     start, bytes);
+			btrfs_add_free_space(block_group, start, bytes);
 			if (!update_ret)
 				btrfs_update_reserved_bytes(block_group,
 							    bytes, 0, 1);
@@ -2342,3 +2416,145 @@ int btrfs_trim_block_group(struct btrfs_block_group_cache *block_group,
 
 	return ret;
 }
+
+/*
+ * Find the left-most item in the cache tree, and then return the
+ * smallest inode number in the item.
+ *
+ * Note: the returned inode number may not be the smallest one in
+ * the tree, if the left-most item is a bitmap.
+ */
+u64 btrfs_find_ino_for_alloc(struct btrfs_root *fs_root)
+{
+	struct btrfs_free_space_ctl *ctl = fs_root->free_ino_ctl;
+	struct btrfs_free_space *entry = NULL;
+	u64 ino = 0;
+
+	spin_lock(&ctl->tree_lock);
+
+	if (RB_EMPTY_ROOT(&ctl->free_space_offset))
+		goto out;
+
+	entry = rb_entry(rb_first(&ctl->free_space_offset),
+			 struct btrfs_free_space, offset_index);
+
+	if (!entry->bitmap) {
+		ino = entry->offset;
+
+		unlink_free_space(ctl, entry);
+		entry->offset++;
+		entry->bytes--;
+		if (!entry->bytes)
+			kmem_cache_free(btrfs_free_space_cachep, entry);
+		else
+			link_free_space(ctl, entry);
+	} else {
+		u64 offset = 0;
+		u64 count = 1;
+		int ret;
+
+		ret = search_bitmap(ctl, entry, &offset, &count);
+		BUG_ON(ret);
+
+		ino = offset;
+		bitmap_clear_bits(ctl, entry, offset, 1);
+		if (entry->bytes == 0)
+			free_bitmap(ctl, entry);
+	}
+out:
+	spin_unlock(&ctl->tree_lock);
+
+	return ino;
+}
+
+struct inode *lookup_free_ino_inode(struct btrfs_root *root,
+				    struct btrfs_path *path)
+{
+	struct inode *inode = NULL;
+
+	spin_lock(&root->cache_lock);
+	if (root->cache_inode)
+		inode = igrab(root->cache_inode);
+	spin_unlock(&root->cache_lock);
+	if (inode)
+		return inode;
+
+	inode = __lookup_free_space_inode(root, path, 0);
+	if (IS_ERR(inode))
+		return inode;
+
+	spin_lock(&root->cache_lock);
+	if (!root->fs_info->closing)
+		root->cache_inode = igrab(inode);
+	spin_unlock(&root->cache_lock);
+
+	return inode;
+}
+
+int create_free_ino_inode(struct btrfs_root *root,
+			  struct btrfs_trans_handle *trans,
+			  struct btrfs_path *path)
+{
+	return __create_free_space_inode(root, trans, path,
+					 BTRFS_FREE_INO_OBJECTID, 0);
+}
+
+int load_free_ino_cache(struct btrfs_fs_info *fs_info, struct btrfs_root *root)
+{
+	struct btrfs_free_space_ctl *ctl = root->free_ino_ctl;
+	struct btrfs_path *path;
+	struct inode *inode;
+	int ret = 0;
+	u64 root_gen = btrfs_root_generation(&root->root_item);
+
+	/*
+	 * If we're unmounting then just return, since this does a search on the
+	 * normal root and not the commit root and we could deadlock.
+	 */
+	smp_mb();
+	if (fs_info->closing)
+		return 0;
+
+	path = btrfs_alloc_path();
+	if (!path)
+		return 0;
+
+	inode = lookup_free_ino_inode(root, path);
+	if (IS_ERR(inode))
+		goto out;
+
+	if (root_gen != BTRFS_I(inode)->generation)
+		goto out_put;
+
+	ret = __load_free_space_cache(root, inode, ctl, path, 0);
+
+	if (ret < 0)
+		printk(KERN_ERR "btrfs: failed to load free ino cache for "
+		       "root %llu\n", root->root_key.objectid);
+out_put:
+	iput(inode);
+out:
+	btrfs_free_path(path);
+	return ret;
+}
+
+int btrfs_write_out_ino_cache(struct btrfs_root *root,
+			      struct btrfs_trans_handle *trans,
+			      struct btrfs_path *path)
+{
+	struct btrfs_free_space_ctl *ctl = root->free_ino_ctl;
+	struct inode *inode;
+	int ret;
+
+	inode = lookup_free_ino_inode(root, path);
+	if (IS_ERR(inode))
+		return 0;
+
+	ret = __btrfs_write_out_cache(root, inode, ctl, NULL, trans, path, 0);
+	if (ret < 0)
+		printk(KERN_ERR "btrfs: failed to write free ino cache "
+		       "for root %llu\n", root->root_key.objectid);
+
+	iput(inode);
+	return ret;
+}

fs/btrfs/free-space-cache.h

@@ -27,6 +27,25 @@ struct btrfs_free_space {
 	struct list_head list;
 };
 
+struct btrfs_free_space_ctl {
+	spinlock_t tree_lock;
+	struct rb_root free_space_offset;
+	u64 free_space;
+	int extents_thresh;
+	int free_extents;
+	int total_bitmaps;
+	int unit;
+	u64 start;
+	struct btrfs_free_space_op *op;
+	void *private;
+};
+
+struct btrfs_free_space_op {
+	void (*recalc_thresholds)(struct btrfs_free_space_ctl *ctl);
+	bool (*use_bitmap)(struct btrfs_free_space_ctl *ctl,
+			   struct btrfs_free_space *info);
+};
+
 struct inode *lookup_free_space_inode(struct btrfs_root *root,
 				      struct btrfs_block_group_cache
 				      *block_group, struct btrfs_path *path);
@@ -45,17 +64,38 @@ int btrfs_write_out_cache(struct btrfs_root *root,
 			  struct btrfs_trans_handle *trans,
 			  struct btrfs_block_group_cache *block_group,
 			  struct btrfs_path *path);
-int btrfs_add_free_space(struct btrfs_block_group_cache *block_group,
-			 u64 bytenr, u64 size);
+
+struct inode *lookup_free_ino_inode(struct btrfs_root *root,
+				    struct btrfs_path *path);
+int create_free_ino_inode(struct btrfs_root *root,
+			  struct btrfs_trans_handle *trans,
+			  struct btrfs_path *path);
+int load_free_ino_cache(struct btrfs_fs_info *fs_info,
+			struct btrfs_root *root);
+int btrfs_write_out_ino_cache(struct btrfs_root *root,
+			      struct btrfs_trans_handle *trans,
+			      struct btrfs_path *path);
+
+void btrfs_init_free_space_ctl(struct btrfs_block_group_cache *block_group);
+int __btrfs_add_free_space(struct btrfs_free_space_ctl *ctl,
+			   u64 bytenr, u64 size);
+static inline int
+btrfs_add_free_space(struct btrfs_block_group_cache *block_group,
+		     u64 bytenr, u64 size)
+{
+	return __btrfs_add_free_space(block_group->free_space_ctl,
+				      bytenr, size);
+}
 int btrfs_remove_free_space(struct btrfs_block_group_cache *block_group,
 			    u64 bytenr, u64 size);
+void __btrfs_remove_free_space_cache(struct btrfs_free_space_ctl *ctl);
 void btrfs_remove_free_space_cache(struct btrfs_block_group_cache
-				   *block_group);
+				     *block_group);
 u64 btrfs_find_space_for_alloc(struct btrfs_block_group_cache *block_group,
 			       u64 offset, u64 bytes, u64 empty_size);
+u64 btrfs_find_ino_for_alloc(struct btrfs_root *fs_root);
 void btrfs_dump_free_space(struct btrfs_block_group_cache *block_group,
 			   u64 bytes);
-u64 btrfs_block_group_free_space(struct btrfs_block_group_cache *block_group);
 int btrfs_find_space_cluster(struct btrfs_trans_handle *trans,
 			     struct btrfs_root *root,
 			     struct btrfs_block_group_cache *block_group,

fs/btrfs/inode-map.c

@@ -16,11 +16,430 @@
  * Boston, MA 021110-1307, USA.
  */
 
+#include <linux/delay.h>
+#include <linux/kthread.h>
+#include <linux/pagemap.h>
+
 #include "ctree.h"
 #include "disk-io.h"
+#include "free-space-cache.h"
+#include "inode-map.h"
 #include "transaction.h"
 
-int btrfs_find_highest_inode(struct btrfs_root *root, u64 *objectid)
+static int caching_kthread(void *data)
+{
+	struct btrfs_root *root = data;
+	struct btrfs_fs_info *fs_info = root->fs_info;
+	struct btrfs_free_space_ctl *ctl = root->free_ino_ctl;
+	struct btrfs_key key;
+	struct btrfs_path *path;
+	struct extent_buffer *leaf;
+	u64 last = (u64)-1;
+	int slot;
+	int ret;
+
+	path = btrfs_alloc_path();
+	if (!path)
+		return -ENOMEM;
+
+	/* Since the commit root is read-only, we can safely skip locking. */
+	path->skip_locking = 1;
+	path->search_commit_root = 1;
+	path->reada = 2;
+
+	key.objectid = BTRFS_FIRST_FREE_OBJECTID;
+	key.offset = 0;
+	key.type = BTRFS_INODE_ITEM_KEY;
+again:
+	/* need to make sure the commit_root doesn't disappear */
+	mutex_lock(&root->fs_commit_mutex);
+
+	ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
+	if (ret < 0)
+		goto out;
+
+	while (1) {
+		smp_mb();
+		if (fs_info->closing > 1)
+			goto out;
+
+		leaf = path->nodes[0];
+		slot = path->slots[0];
+		if (path->slots[0] >= btrfs_header_nritems(leaf)) {
+			ret = btrfs_next_leaf(root, path);
+			if (ret < 0)
+				goto out;
+			else if (ret > 0)
+				break;
+
+			if (need_resched() ||
+			    btrfs_transaction_in_commit(fs_info)) {
+				leaf = path->nodes[0];
+
+				if (btrfs_header_nritems(leaf) == 0) {
+					WARN_ON(1);
+					break;
+				}
+
+				/*
+				 * Save the key so we can advances forward
+				 * in the next search.
+				 */
+				btrfs_item_key_to_cpu(leaf, &key, 0);
+				btrfs_release_path(root, path);
+				root->cache_progress = last;
+				mutex_unlock(&root->fs_commit_mutex);
+				schedule_timeout(1);
+				goto again;
+			} else
+				continue;
+		}
+
+		btrfs_item_key_to_cpu(leaf, &key, slot);
+
+		if (key.type != BTRFS_INODE_ITEM_KEY)
+			goto next;
+
+		if (key.objectid >= BTRFS_LAST_FREE_OBJECTID)
+			break;
+
+		if (last != (u64)-1 && last + 1 != key.objectid) {
+			__btrfs_add_free_space(ctl, last + 1,
+					       key.objectid - last - 1);
+			wake_up(&root->cache_wait);
+		}
+
+		last = key.objectid;
+next:
+		path->slots[0]++;
+	}
+
+	if (last < BTRFS_LAST_FREE_OBJECTID - 1) {
+		__btrfs_add_free_space(ctl, last + 1,
+				       BTRFS_LAST_FREE_OBJECTID - last - 1);
+	}
+
+	spin_lock(&root->cache_lock);
+	root->cached = BTRFS_CACHE_FINISHED;
+	spin_unlock(&root->cache_lock);
+
+	root->cache_progress = (u64)-1;
+	btrfs_unpin_free_ino(root);
+out:
+	wake_up(&root->cache_wait);
+	mutex_unlock(&root->fs_commit_mutex);
+
+	btrfs_free_path(path);
+
+	return ret;
+}
+
+static void start_caching(struct btrfs_root *root)
+{
+	struct task_struct *tsk;
+	int ret;
+
+	spin_lock(&root->cache_lock);
+	if (root->cached != BTRFS_CACHE_NO) {
+		spin_unlock(&root->cache_lock);
+		return;
+	}
+
+	root->cached = BTRFS_CACHE_STARTED;
+	spin_unlock(&root->cache_lock);
+
+	ret = load_free_ino_cache(root->fs_info, root);
+	if (ret == 1) {
+		spin_lock(&root->cache_lock);
+		root->cached = BTRFS_CACHE_FINISHED;
+		spin_unlock(&root->cache_lock);
+		return;
+	}
+
+	tsk = kthread_run(caching_kthread, root, "btrfs-ino-cache-%llu\n",
+			  root->root_key.objectid);
+	BUG_ON(IS_ERR(tsk));
+}
+
+int btrfs_find_free_ino(struct btrfs_root *root, u64 *objectid)
+{
+again:
+	*objectid = btrfs_find_ino_for_alloc(root);
+
+	if (*objectid != 0)
+		return 0;
+
+	start_caching(root);
+
+	wait_event(root->cache_wait,
+		   root->cached == BTRFS_CACHE_FINISHED ||
+		   root->free_ino_ctl->free_space > 0);
+
+	if (root->cached == BTRFS_CACHE_FINISHED &&
+	    root->free_ino_ctl->free_space == 0)
+		return -ENOSPC;
+	else
+		goto again;
+}
+
+void btrfs_return_ino(struct btrfs_root *root, u64 objectid)
+{
+	struct btrfs_free_space_ctl *ctl = root->free_ino_ctl;
+	struct btrfs_free_space_ctl *pinned = root->free_ino_pinned;
+again:
+	if (root->cached == BTRFS_CACHE_FINISHED) {
+		__btrfs_add_free_space(ctl, objectid, 1);
+	} else {
+		/*
+		 * If we are in the process of caching free ino chunks,
+		 * to avoid adding the same inode number to the free_ino
+		 * tree twice due to cross transaction, we'll leave it
+		 * in the pinned tree until a transaction is committed
+		 * or the caching work is done.
+		 */
+
+		mutex_lock(&root->fs_commit_mutex);
+		spin_lock(&root->cache_lock);
+		if (root->cached == BTRFS_CACHE_FINISHED) {
+			spin_unlock(&root->cache_lock);
+			mutex_unlock(&root->fs_commit_mutex);
+			goto again;
+		}
+		spin_unlock(&root->cache_lock);
+
+		start_caching(root);
+
+		if (objectid <= root->cache_progress)
+			__btrfs_add_free_space(ctl, objectid, 1);
+		else
+			__btrfs_add_free_space(pinned, objectid, 1);
+
+		mutex_unlock(&root->fs_commit_mutex);
+	}
+}
+
+/*
+ * When a transaction is committed, we'll move those inode numbers which
+ * are smaller than root->cache_progress from pinned tree to free_ino tree,
+ * and others will just be dropped, because the commit root we were
+ * searching has changed.
+ *
+ * Must be called with root->fs_commit_mutex held
+ */
+void btrfs_unpin_free_ino(struct btrfs_root *root)
+{
+	struct btrfs_free_space_ctl *ctl = root->free_ino_ctl;
+	struct rb_root *rbroot = &root->free_ino_pinned->free_space_offset;
+	struct btrfs_free_space *info;
+	struct rb_node *n;
+	u64 count;
+
+	while (1) {
+		n = rb_first(rbroot);
+		if (!n)
+			break;
+
+		info = rb_entry(n, struct btrfs_free_space, offset_index);
+		BUG_ON(info->bitmap);
+
+		if (info->offset > root->cache_progress)
+			goto free;
+		else if (info->offset + info->bytes > root->cache_progress)
+			count = root->cache_progress - info->offset + 1;
+		else
+			count = info->bytes;
+
+		__btrfs_add_free_space(ctl, info->offset, count);
+free:
+		rb_erase(&info->offset_index, rbroot);
+		kfree(info);
+	}
+}
+
+#define INIT_THRESHOLD	(((1024 * 32) / 2) / sizeof(struct btrfs_free_space))
+#define INODES_PER_BITMAP (PAGE_CACHE_SIZE * 8)
+
+/*
+ * The goal is to keep the memory used by the free_ino tree won't
+ * exceed the memory if we use bitmaps only.
+ */
+static void recalculate_thresholds(struct btrfs_free_space_ctl *ctl)
+{
+	struct btrfs_free_space *info;
+	struct rb_node *n;
+	int max_ino;
+	int max_bitmaps;
+
+	n = rb_last(&ctl->free_space_offset);
+	if (!n) {
+		ctl->extents_thresh = INIT_THRESHOLD;
+		return;
+	}
+	info = rb_entry(n, struct btrfs_free_space, offset_index);
+
+	/*
+	 * Find the maximum inode number in the filesystem. Note we
+	 * ignore the fact that this can be a bitmap, because we are
+	 * not doing precise calculation.
+	 */
+	max_ino = info->bytes - 1;
+
+	max_bitmaps = ALIGN(max_ino, INODES_PER_BITMAP) / INODES_PER_BITMAP;
+	if (max_bitmaps <= ctl->total_bitmaps) {
+		ctl->extents_thresh = 0;
+		return;
+	}
+
+	ctl->extents_thresh = (max_bitmaps - ctl->total_bitmaps) *
+				PAGE_CACHE_SIZE / sizeof(*info);
+}
+
+/*
+ * We don't fall back to bitmap, if we are below the extents threshold
+ * or this chunk of inode numbers is a big one.
+ */
+static bool use_bitmap(struct btrfs_free_space_ctl *ctl,
+		       struct btrfs_free_space *info)
+{
+	if (ctl->free_extents < ctl->extents_thresh ||
+	    info->bytes > INODES_PER_BITMAP / 10)
+		return false;
+
+	return true;
+}
+
+static struct btrfs_free_space_op free_ino_op = {
+	.recalc_thresholds	= recalculate_thresholds,
+	.use_bitmap		= use_bitmap,
+};
+
+static void pinned_recalc_thresholds(struct btrfs_free_space_ctl *ctl)
+{
+}
+
+static bool pinned_use_bitmap(struct btrfs_free_space_ctl *ctl,
+			      struct btrfs_free_space *info)
+{
+	/*
+	 * We always use extents for two reasons:
+	 *
+	 * - The pinned tree is only used during the process of caching
+	 *   work.
+	 * - Make code simpler. See btrfs_unpin_free_ino().
+	 */
+	return false;
+}
+
+static struct btrfs_free_space_op pinned_free_ino_op = {
+	.recalc_thresholds	= pinned_recalc_thresholds,
+	.use_bitmap		= pinned_use_bitmap,
+};
+
+void btrfs_init_free_ino_ctl(struct btrfs_root *root)
+{
+	struct btrfs_free_space_ctl *ctl = root->free_ino_ctl;
+	struct btrfs_free_space_ctl *pinned = root->free_ino_pinned;
+
+	spin_lock_init(&ctl->tree_lock);
+	ctl->unit = 1;
+	ctl->start = 0;
+	ctl->private = NULL;
+	ctl->op = &free_ino_op;
+
+	/*
+	 * Initially we allow to use 16K of ram to cache chunks of
+	 * inode numbers before we resort to bitmaps. This is somewhat
+	 * arbitrary, but it will be adjusted in runtime.
+	 */
+	ctl->extents_thresh = INIT_THRESHOLD;
+
+	spin_lock_init(&pinned->tree_lock);
+	pinned->unit = 1;
+	pinned->start = 0;
+	pinned->private = NULL;
+	pinned->extents_thresh = 0;
+	pinned->op = &pinned_free_ino_op;
+}
+
+int btrfs_save_ino_cache(struct btrfs_root *root,
+			 struct btrfs_trans_handle *trans)
+{
+	struct btrfs_free_space_ctl *ctl = root->free_ino_ctl;
+	struct btrfs_path *path;
+	struct inode *inode;
+	u64 alloc_hint = 0;
+	int ret;
+	int prealloc;
+	bool retry = false;
+
+	path = btrfs_alloc_path();
+	if (!path)
+		return -ENOMEM;
+again:
+	inode = lookup_free_ino_inode(root, path);
+	if (IS_ERR(inode) && PTR_ERR(inode) != -ENOENT) {
+		ret = PTR_ERR(inode);
+		goto out;
+	}
+
+	if (IS_ERR(inode)) {
+		BUG_ON(retry);
+		retry = true;
+
+		ret = create_free_ino_inode(root, trans, path);
+		if (ret)
+			goto out;
+		goto again;
+	}
+
+	BTRFS_I(inode)->generation = 0;
+	ret = btrfs_update_inode(trans, root, inode);
+	WARN_ON(ret);
+
+	if (i_size_read(inode) > 0) {
+		ret = btrfs_truncate_free_space_cache(root, trans, path, inode);
+		if (ret)
+			goto out_put;
+	}
+
+	spin_lock(&root->cache_lock);
+	if (root->cached != BTRFS_CACHE_FINISHED) {
+		ret = -1;
+		spin_unlock(&root->cache_lock);
+		goto out_put;
+	}
+	spin_unlock(&root->cache_lock);
+
+	spin_lock(&ctl->tree_lock);
+	prealloc = sizeof(struct btrfs_free_space) * ctl->free_extents;
+	prealloc = ALIGN(prealloc, PAGE_CACHE_SIZE);
+	prealloc += ctl->total_bitmaps * PAGE_CACHE_SIZE;
+	spin_unlock(&ctl->tree_lock);
+
+	/* Just to make sure we have enough space */
+	prealloc += 8 * PAGE_CACHE_SIZE;
+
+	ret = btrfs_check_data_free_space(inode, prealloc);
+	if (ret)
+		goto out_put;
+
+	ret = btrfs_prealloc_file_range_trans(inode, trans, 0, 0, prealloc,
+					      prealloc, prealloc, &alloc_hint);
+	if (ret)
+		goto out_put;
+	btrfs_free_reserved_data_space(inode, prealloc);
+
+out_put:
+	iput(inode);
+out:
+	if (ret == 0)
+		ret = btrfs_write_out_ino_cache(root, trans, path);
+
+	btrfs_free_path(path);
+	return ret;
+}
+
+static int btrfs_find_highest_objectid(struct btrfs_root *root, u64 *objectid)
 {
 	struct btrfs_path *path;
 	int ret;
@@ -55,15 +474,14 @@ error:
 	return ret;
 }
 
-int btrfs_find_free_objectid(struct btrfs_trans_handle *trans,
-			     struct btrfs_root *root,
-			     u64 dirid, u64 *objectid)
+int btrfs_find_free_objectid(struct btrfs_root *root, u64 *objectid)
 {
 	int ret;
 	mutex_lock(&root->objectid_mutex);
 
 	if (unlikely(root->highest_objectid < BTRFS_FIRST_FREE_OBJECTID)) {
-		ret = btrfs_find_highest_inode(root, &root->highest_objectid);
+		ret = btrfs_find_highest_objectid(root,
+						  &root->highest_objectid);
 		if (ret)
 			goto out;
 	}

fs/btrfs/inode-map.h

@@ -0,0 +1,13 @@
+#ifndef __BTRFS_INODE_MAP
+#define __BTRFS_INODE_MAP
+
+void btrfs_init_free_ino_ctl(struct btrfs_root *root);
+void btrfs_unpin_free_ino(struct btrfs_root *root);
+void btrfs_return_ino(struct btrfs_root *root, u64 objectid);
+int btrfs_find_free_ino(struct btrfs_root *root, u64 *objectid);
+int btrfs_save_ino_cache(struct btrfs_root *root,
+			 struct btrfs_trans_handle *trans);
+
+int btrfs_find_free_objectid(struct btrfs_root *root, u64 *objectid);
+
+#endif

fs/btrfs/inode.c

@@ -51,6 +51,7 @@
 #include "compression.h"
 #include "locking.h"
 #include "free-space-cache.h"
+#include "inode-map.h"
 
 struct btrfs_iget_args {
 	u64 ino;
@@ -138,7 +139,7 @@ static noinline int insert_inline_extent(struct btrfs_trans_handle *trans,
 	path->leave_spinning = 1;
 	btrfs_set_trans_block_group(trans, inode);
 
-	key.objectid = inode->i_ino;
+	key.objectid = btrfs_ino(inode);
 	key.offset = start;
 	btrfs_set_key_type(&key, BTRFS_EXTENT_DATA_KEY);
 	datasize = btrfs_file_extent_calc_inline_size(cur_size);
@@ -745,6 +746,15 @@ static u64 get_extent_allocation_hint(struct inode *inode, u64 start,
 	return alloc_hint;
 }
 
+static inline bool is_free_space_inode(struct btrfs_root *root,
+				       struct inode *inode)
+{
+	if (root == root->fs_info->tree_root ||
+	    BTRFS_I(inode)->location.objectid == BTRFS_FREE_INO_OBJECTID)
+		return true;
+	return false;
+}
+
 /*
  * when extent_io.c finds a delayed allocation range in the file,
  * the call backs end up in this code.  The basic idea is to
@@ -777,7 +787,7 @@ static noinline int cow_file_range(struct inode *inode,
 	struct extent_map_tree *em_tree = &BTRFS_I(inode)->extent_tree;
 	int ret = 0;
 
-	BUG_ON(root == root->fs_info->tree_root);
+	BUG_ON(is_free_space_inode(root, inode));
 	trans = btrfs_join_transaction(root, 1);
 	BUG_ON(IS_ERR(trans));
 	btrfs_set_trans_block_group(trans, inode);
@@ -1049,29 +1059,31 @@ static noinline int run_delalloc_nocow(struct inode *inode,
 	int type;
 	int nocow;
 	int check_prev = 1;
-	bool nolock = false;
+	bool nolock;
+	u64 ino = btrfs_ino(inode);
 
 	path = btrfs_alloc_path();
 	BUG_ON(!path);
-	if (root == root->fs_info->tree_root) {
-		nolock = true;
+
+	nolock = is_free_space_inode(root, inode);
+
+	if (nolock)
 		trans = btrfs_join_transaction_nolock(root, 1);
-	} else {
+	else
 		trans = btrfs_join_transaction(root, 1);
-	}
 	BUG_ON(IS_ERR(trans));
 
 	cow_start = (u64)-1;
 	cur_offset = start;
 	while (1) {
-		ret = btrfs_lookup_file_extent(trans, root, path, inode->i_ino,
+		ret = btrfs_lookup_file_extent(trans, root, path, ino,
 					       cur_offset, 0);
 		BUG_ON(ret < 0);
 		if (ret > 0 && path->slots[0] > 0 && check_prev) {
 			leaf = path->nodes[0];
 			btrfs_item_key_to_cpu(leaf, &found_key,
 					      path->slots[0] - 1);
-			if (found_key.objectid == inode->i_ino &&
+			if (found_key.objectid == ino &&
 			    found_key.type == BTRFS_EXTENT_DATA_KEY)
 				path->slots[0]--;
 		}
@@ -1092,7 +1104,7 @@ next_slot:
 		num_bytes = 0;
 		btrfs_item_key_to_cpu(leaf, &found_key, path->slots[0]);
 
-		if (found_key.objectid > inode->i_ino ||
+		if (found_key.objectid > ino ||
 		    found_key.type > BTRFS_EXTENT_DATA_KEY ||
 		    found_key.offset > end)
 			break;
@@ -1127,7 +1139,7 @@ next_slot:
 				goto out_check;
 			if (btrfs_extent_readonly(root, disk_bytenr))
 				goto out_check;
-			if (btrfs_cross_ref_exist(trans, root, inode->i_ino,
+			if (btrfs_cross_ref_exist(trans, root, ino,
 						  found_key.offset -
 						  extent_offset, disk_bytenr))
 				goto out_check;
@@ -1316,8 +1328,7 @@ static int btrfs_set_bit_hook(struct inode *inode,
 	if (!(state->state & EXTENT_DELALLOC) && (*bits & EXTENT_DELALLOC)) {
 		struct btrfs_root *root = BTRFS_I(inode)->root;
 		u64 len = state->end + 1 - state->start;
-		int do_list = (root->root_key.objectid !=
-			       BTRFS_ROOT_TREE_OBJECTID);
+		bool do_list = !is_free_space_inode(root, inode);
 
 		if (*bits & EXTENT_FIRST_DELALLOC)
 			*bits &= ~EXTENT_FIRST_DELALLOC;
@@ -1350,8 +1361,7 @@ static int btrfs_clear_bit_hook(struct inode *inode,
 	if ((state->state & EXTENT_DELALLOC) && (*bits & EXTENT_DELALLOC)) {
 		struct btrfs_root *root = BTRFS_I(inode)->root;
 		u64 len = state->end + 1 - state->start;
-		int do_list = (root->root_key.objectid !=
-			       BTRFS_ROOT_TREE_OBJECTID);
+		bool do_list = !is_free_space_inode(root, inode);
 
 		if (*bits & EXTENT_FIRST_DELALLOC)
 			*bits &= ~EXTENT_FIRST_DELALLOC;
@@ -1458,7 +1468,7 @@ static int btrfs_submit_bio_hook(struct inode *inode, int rw, struct bio *bio,
 
 	skip_sum = BTRFS_I(inode)->flags & BTRFS_INODE_NODATASUM;
 
-	if (root == root->fs_info->tree_root)
+	if (is_free_space_inode(root, inode))
 		ret = btrfs_bio_wq_end_io(root->fs_info, bio, 2);
 	else
 		ret = btrfs_bio_wq_end_io(root->fs_info, bio, 0);
@@ -1644,7 +1654,7 @@ static int insert_reserved_file_extent(struct btrfs_trans_handle *trans,
 				 &hint, 0);
 	BUG_ON(ret);
 
-	ins.objectid = inode->i_ino;
+	ins.objectid = btrfs_ino(inode);
 	ins.offset = file_pos;
 	ins.type = BTRFS_EXTENT_DATA_KEY;
 	ret = btrfs_insert_empty_item(trans, root, path, &ins, sizeof(*fi));
@@ -1675,7 +1685,7 @@ static int insert_reserved_file_extent(struct btrfs_trans_handle *trans,
 	ins.type = BTRFS_EXTENT_ITEM_KEY;
 	ret = btrfs_alloc_reserved_file_extent(trans, root,
 					root->root_key.objectid,
-					inode->i_ino, file_pos, &ins);
+					btrfs_ino(inode), file_pos, &ins);
 	BUG_ON(ret);
 	btrfs_free_path(path);
 
@@ -1701,7 +1711,7 @@ static int btrfs_finish_ordered_io(struct inode *inode, u64 start, u64 end)
 	struct extent_state *cached_state = NULL;
 	int compress_type = 0;
 	int ret;
-	bool nolock = false;
+	bool nolock;
 
 	ret = btrfs_dec_test_ordered_pending(inode, &ordered_extent, start,
 					     end - start + 1);
@@ -1709,7 +1719,7 @@ static int btrfs_finish_ordered_io(struct inode *inode, u64 start, u64 end)
 		return 0;
 	BUG_ON(!ordered_extent);
 
-	nolock = (root == root->fs_info->tree_root);
+	nolock = is_free_space_inode(root, inode);
 
 	if (test_bit(BTRFS_ORDERED_NOCOW, &ordered_extent->flags)) {
 		BUG_ON(!list_empty(&ordered_extent->list));
@@ -2005,8 +2015,9 @@ good:
 
 zeroit:
 	if (printk_ratelimit()) {
-		printk(KERN_INFO "btrfs csum failed ino %lu off %llu csum %u "
-		       "private %llu\n", page->mapping->host->i_ino,
+		printk(KERN_INFO "btrfs csum failed ino %llu off %llu csum %u "
+		       "private %llu\n",
+		       (unsigned long long)btrfs_ino(page->mapping->host),
 		       (unsigned long long)start, csum,
 		       (unsigned long long)private);
 	}
@@ -2244,7 +2255,7 @@ int btrfs_orphan_add(struct btrfs_trans_handle *trans, struct inode *inode)
 
 	/* insert an orphan item to track this unlinked/truncated file */
 	if (insert >= 1) {
-		ret = btrfs_insert_orphan_item(trans, root, inode->i_ino);
+		ret = btrfs_insert_orphan_item(trans, root, btrfs_ino(inode));
 		BUG_ON(ret);
 	}
 
@@ -2281,7 +2292,7 @@ int btrfs_orphan_del(struct btrfs_trans_handle *trans, struct inode *inode)
 	spin_unlock(&root->orphan_lock);
 
 	if (trans && delete_item) {
-		ret = btrfs_del_orphan_item(trans, root, inode->i_ino);
+		ret = btrfs_del_orphan_item(trans, root, btrfs_ino(inode));
 		BUG_ON(ret);
 	}
 
@@ -2543,7 +2554,8 @@ static void btrfs_read_locked_inode(struct inode *inode)
 	 * try to precache a NULL acl entry for files that don't have
 	 * any xattrs or acls
 	 */
-	maybe_acls = acls_after_inode_item(leaf, path->slots[0], inode->i_ino);
+	maybe_acls = acls_after_inode_item(leaf, path->slots[0],
+					   btrfs_ino(inode));
 	if (!maybe_acls)
 		cache_no_acl(inode);
 
@@ -2689,6 +2701,8 @@ static int __btrfs_unlink_inode(struct btrfs_trans_handle *trans,
 	struct btrfs_dir_item *di;
 	struct btrfs_key key;
 	u64 index;
+	u64 ino = btrfs_ino(inode);
+	u64 dir_ino = btrfs_ino(dir);
 
 	path = btrfs_alloc_path();
 	if (!path) {
@@ -2697,7 +2711,7 @@ static int __btrfs_unlink_inode(struct btrfs_trans_handle *trans,
 	}
 
 	path->leave_spinning = 1;
-	di = btrfs_lookup_dir_item(trans, root, path, dir->i_ino,
+	di = btrfs_lookup_dir_item(trans, root, path, dir_ino,
 				    name, name_len, -1);
 	if (IS_ERR(di)) {
 		ret = PTR_ERR(di);
@@ -2714,17 +2728,16 @@ static int __btrfs_unlink_inode(struct btrfs_trans_handle *trans,
 		goto err;
 	btrfs_release_path(root, path);
 
-	ret = btrfs_del_inode_ref(trans, root, name, name_len,
-				  inode->i_ino,
-				  dir->i_ino, &index);
+	ret = btrfs_del_inode_ref(trans, root, name, name_len, ino,
+				  dir_ino, &index);
 	if (ret) {
 		printk(KERN_INFO "btrfs failed to delete reference to %.*s, "
-		       "inode %lu parent %lu\n", name_len, name,
-		       inode->i_ino, dir->i_ino);
+		       "inode %llu parent %llu\n", name_len, name,
+		       (unsigned long long)ino, (unsigned long long)dir_ino);
 		goto err;
 	}
 
-	di = btrfs_lookup_dir_index_item(trans, root, path, dir->i_ino,
+	di = btrfs_lookup_dir_index_item(trans, root, path, dir_ino,
 					 index, name, name_len, -1);
 	if (IS_ERR(di)) {
 		ret = PTR_ERR(di);
@@ -2738,7 +2751,7 @@ static int __btrfs_unlink_inode(struct btrfs_trans_handle *trans,
 	btrfs_release_path(root, path);
 
 	ret = btrfs_del_inode_ref_in_log(trans, root, name, name_len,
-					 inode, dir->i_ino);
+					 inode, dir_ino);
 	BUG_ON(ret != 0 && ret != -ENOENT);
 
 	ret = btrfs_del_dir_entries_in_log(trans, root, name, name_len,
@@ -2816,12 +2829,14 @@ static struct btrfs_trans_handle *__unlink_start_trans(struct inode *dir,
 	int check_link = 1;
 	int err = -ENOSPC;
 	int ret;
+	u64 ino = btrfs_ino(inode);
+	u64 dir_ino = btrfs_ino(dir);
 
 	trans = btrfs_start_transaction(root, 10);
 	if (!IS_ERR(trans) || PTR_ERR(trans) != -ENOSPC)
 		return trans;
 
-	if (inode->i_ino == BTRFS_EMPTY_SUBVOL_DIR_OBJECTID)
+	if (ino == BTRFS_EMPTY_SUBVOL_DIR_OBJECTID)
 		return ERR_PTR(-ENOSPC);
 
 	/* check if there is someone else holds reference */
@@ -2880,7 +2895,7 @@ static struct btrfs_trans_handle *__unlink_start_trans(struct inode *dir,
 
 	if (ret == 0 && S_ISREG(inode->i_mode)) {
 		ret = btrfs_lookup_file_extent(trans, root, path,
-					       inode->i_ino, (u64)-1, 0);
+					       ino, (u64)-1, 0);
 		if (ret < 0) {
 			err = ret;
 			goto out;
@@ -2896,7 +2911,7 @@ static struct btrfs_trans_handle *__unlink_start_trans(struct inode *dir,
 		goto out;
 	}
 
-	di = btrfs_lookup_dir_item(trans, root, path, dir->i_ino,
+	di = btrfs_lookup_dir_item(trans, root, path, dir_ino,
 				dentry->d_name.name, dentry->d_name.len, 0);
 	if (IS_ERR(di)) {
 		err = PTR_ERR(di);
@@ -2913,7 +2928,7 @@ static struct btrfs_trans_handle *__unlink_start_trans(struct inode *dir,
 
 	ref = btrfs_lookup_inode_ref(trans, root, path,
 				dentry->d_name.name, dentry->d_name.len,
-				inode->i_ino, dir->i_ino, 0);
+				ino, dir_ino, 0);
 	if (IS_ERR(ref)) {
 		err = PTR_ERR(ref);
 		goto out;
@@ -2924,7 +2939,7 @@ static struct btrfs_trans_handle *__unlink_start_trans(struct inode *dir,
 	index = btrfs_inode_ref_index(path->nodes[0], ref);
 	btrfs_release_path(root, path);
 
-	di = btrfs_lookup_dir_index_item(trans, root, path, dir->i_ino, index,
+	di = btrfs_lookup_dir_index_item(trans, root, path, dir_ino, index,
 				dentry->d_name.name, dentry->d_name.len, 0);
 	if (IS_ERR(di)) {
 		err = PTR_ERR(di);
@@ -2999,12 +3014,13 @@ int btrfs_unlink_subvol(struct btrfs_trans_handle *trans,
 	struct btrfs_key key;
 	u64 index;
 	int ret;
+	u64 dir_ino = btrfs_ino(dir);
 
 	path = btrfs_alloc_path();
 	if (!path)
 		return -ENOMEM;
 
-	di = btrfs_lookup_dir_item(trans, root, path, dir->i_ino,
+	di = btrfs_lookup_dir_item(trans, root, path, dir_ino,
 				   name, name_len, -1);
 	BUG_ON(!di || IS_ERR(di));
 
@@ -3017,10 +3033,10 @@ int btrfs_unlink_subvol(struct btrfs_trans_handle *trans,
 
 	ret = btrfs_del_root_ref(trans, root->fs_info->tree_root,
 				 objectid, root->root_key.objectid,
-				 dir->i_ino, &index, name, name_len);
+				 dir_ino, &index, name, name_len);
 	if (ret < 0) {
 		BUG_ON(ret != -ENOENT);
-		di = btrfs_search_dir_index_item(root, path, dir->i_ino,
+		di = btrfs_search_dir_index_item(root, path, dir_ino,
 						 name, name_len);
 		BUG_ON(!di || IS_ERR(di));
 
@@ -3030,7 +3046,7 @@ int btrfs_unlink_subvol(struct btrfs_trans_handle *trans,
 		index = key.offset;
 	}
 
-	di = btrfs_lookup_dir_index_item(trans, root, path, dir->i_ino,
+	di = btrfs_lookup_dir_index_item(trans, root, path, dir_ino,
 					 index, name, name_len, -1);
 	BUG_ON(!di || IS_ERR(di));
 
@@ -3059,7 +3075,7 @@ static int btrfs_rmdir(struct inode *dir, struct dentry *dentry)
 	unsigned long nr = 0;
 
 	if (inode->i_size > BTRFS_EMPTY_DIR_SIZE ||
-	    inode->i_ino == BTRFS_FIRST_FREE_OBJECTID)
+	    btrfs_ino(inode) == BTRFS_FIRST_FREE_OBJECTID)
 		return -ENOTEMPTY;
 
 	trans = __unlink_start_trans(dir, dentry);
@@ -3068,7 +3084,7 @@ static int btrfs_rmdir(struct inode *dir, struct dentry *dentry)
 
 	btrfs_set_trans_block_group(trans, dir);
 
-	if (unlikely(inode->i_ino == BTRFS_EMPTY_SUBVOL_DIR_OBJECTID)) {
+	if (unlikely(btrfs_ino(inode) == BTRFS_EMPTY_SUBVOL_DIR_OBJECTID)) {
 		err = btrfs_unlink_subvol(trans, root, dir,
 					  BTRFS_I(inode)->location.objectid,
 					  dentry->d_name.name,
@@ -3300,6 +3316,7 @@ int btrfs_truncate_inode_items(struct btrfs_trans_handle *trans,
 	int encoding;
 	int ret;
 	int err = 0;
+	u64 ino = btrfs_ino(inode);
 
 	BUG_ON(new_size > 0 && min_type != BTRFS_EXTENT_DATA_KEY);
 
@@ -3310,7 +3327,7 @@ int btrfs_truncate_inode_items(struct btrfs_trans_handle *trans,
 	BUG_ON(!path);
 	path->reada = -1;
 
-	key.objectid = inode->i_ino;
+	key.objectid = ino;
 	key.offset = (u64)-1;
 	key.type = (u8)-1;
 
@@ -3338,7 +3355,7 @@ search_again:
 		found_type = btrfs_key_type(&found_key);
 		encoding = 0;
 
-		if (found_key.objectid != inode->i_ino)
+		if (found_key.objectid != ino)
 			break;
 
 		if (found_type < min_type)
@@ -3457,7 +3474,7 @@ delete:
 			ret = btrfs_free_extent(trans, root, extent_start,
 						extent_num_bytes, 0,
 						btrfs_header_owner(leaf),
-						inode->i_ino, extent_offset);
+						ino, extent_offset);
 			BUG_ON(ret);
 		}
 
@@ -3466,7 +3483,9 @@ delete:
 
 		if (path->slots[0] == 0 ||
 		    path->slots[0] != pending_del_slot) {
-			if (root->ref_cows) {
+			if (root->ref_cows &&
+			    BTRFS_I(inode)->location.objectid !=
+						BTRFS_FREE_INO_OBJECTID) {
 				err = -EAGAIN;
 				goto out;
 			}
@@ -3656,7 +3675,7 @@ int btrfs_cont_expand(struct inode *inode, loff_t oldsize, loff_t size)
 				break;
 
 			err = btrfs_insert_file_extent(trans, root,
-					inode->i_ino, cur_offset, 0,
+					btrfs_ino(inode), cur_offset, 0,
 					0, hole_size, 0, hole_size,
 					0, 0, 0);
 			if (err)
@@ -3758,7 +3777,7 @@ void btrfs_evict_inode(struct inode *inode)
 
 	truncate_inode_pages(&inode->i_data, 0);
 	if (inode->i_nlink && (btrfs_root_refs(&root->root_item) != 0 ||
-			       root == root->fs_info->tree_root))
+			       is_free_space_inode(root, inode)))
 		goto no_delete;
 
 	if (is_bad_inode(inode)) {
@@ -3811,6 +3830,10 @@ void btrfs_evict_inode(struct inode *inode)
 		BUG_ON(ret);
 	}
 
+	if (!(root == root->fs_info->tree_root ||
+	      root->root_key.objectid == BTRFS_TREE_RELOC_OBJECTID))
+		btrfs_return_ino(root, btrfs_ino(inode));
+
 	nr = trans->blocks_used;
 	btrfs_end_transaction(trans, root);
 	btrfs_btree_balance_dirty(root, nr);
@@ -3836,7 +3859,7 @@ static int btrfs_inode_by_name(struct inode *dir, struct dentry *dentry,
 	path = btrfs_alloc_path();
 	BUG_ON(!path);
 
-	di = btrfs_lookup_dir_item(NULL, root, path, dir->i_ino, name,
+	di = btrfs_lookup_dir_item(NULL, root, path, btrfs_ino(dir), name,
 				    namelen, 0);
 	if (IS_ERR(di))
 		ret = PTR_ERR(di);
@@ -3889,7 +3912,7 @@ static int fixup_tree_root_location(struct btrfs_root *root,
 
 	leaf = path->nodes[0];
 	ref = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_root_ref);
-	if (btrfs_root_ref_dirid(leaf, ref) != dir->i_ino ||
+	if (btrfs_root_ref_dirid(leaf, ref) != btrfs_ino(dir) ||
 	    btrfs_root_ref_name_len(leaf, ref) != dentry->d_name.len)
 		goto out;
 
@@ -3928,6 +3951,7 @@ static void inode_tree_add(struct inode *inode)
 	struct btrfs_inode *entry;
 	struct rb_node **p;
 	struct rb_node *parent;
+	u64 ino = btrfs_ino(inode);
 again:
 	p = &root->inode_tree.rb_node;
 	parent = NULL;
@@ -3940,9 +3964,9 @@ again:
 		parent = *p;
 		entry = rb_entry(parent, struct btrfs_inode, rb_node);
 
-		if (inode->i_ino < entry->vfs_inode.i_ino)
+		if (ino < btrfs_ino(&entry->vfs_inode))
 			p = &parent->rb_left;
-		else if (inode->i_ino > entry->vfs_inode.i_ino)
+		else if (ino > btrfs_ino(&entry->vfs_inode))
 			p = &parent->rb_right;
 		else {
 			WARN_ON(!(entry->vfs_inode.i_state &
@@ -4006,9 +4030,9 @@ again:
 		prev = node;
 		entry = rb_entry(node, struct btrfs_inode, rb_node);
 
-		if (objectid < entry->vfs_inode.i_ino)
+		if (objectid < btrfs_ino(&entry->vfs_inode))
 			node = node->rb_left;
-		else if (objectid > entry->vfs_inode.i_ino)
+		else if (objectid > btrfs_ino(&entry->vfs_inode))
 			node = node->rb_right;
 		else
 			break;
@@ -4016,7 +4040,7 @@ again:
 	if (!node) {
 		while (prev) {
 			entry = rb_entry(prev, struct btrfs_inode, rb_node);
-			if (objectid <= entry->vfs_inode.i_ino) {
+			if (objectid <= btrfs_ino(&entry->vfs_inode)) {
 				node = prev;
 				break;
 			}
@@ -4025,7 +4049,7 @@ again:
 	}
 	while (node) {
 		entry = rb_entry(node, struct btrfs_inode, rb_node);
-		objectid = entry->vfs_inode.i_ino + 1;
+		objectid = btrfs_ino(&entry->vfs_inode) + 1;
 		inode = igrab(&entry->vfs_inode);
 		if (inode) {
 			spin_unlock(&root->inode_lock);
@@ -4063,7 +4087,7 @@ static int btrfs_init_locked_inode(struct inode *inode, void *p)
 static int btrfs_find_actor(struct inode *inode, void *opaque)
 {
 	struct btrfs_iget_args *args = opaque;
-	return args->ino == inode->i_ino &&
+	return args->ino == btrfs_ino(inode) &&
 		args->root == BTRFS_I(inode)->root;
 }
 
@@ -4241,9 +4265,7 @@ static int btrfs_real_readdir(struct file *filp, void *dirent,
 
 	/* special case for "." */
 	if (filp->f_pos == 0) {
-		over = filldir(dirent, ".", 1,
-			       1, inode->i_ino,
-			       DT_DIR);
+		over = filldir(dirent, ".", 1, 1, btrfs_ino(inode), DT_DIR);
 		if (over)
 			return 0;
 		filp->f_pos = 1;
@@ -4262,7 +4284,7 @@ static int btrfs_real_readdir(struct file *filp, void *dirent,
 
 	btrfs_set_key_type(&key, key_type);
 	key.offset = filp->f_pos;
-	key.objectid = inode->i_ino;
+	key.objectid = btrfs_ino(inode);
 
 	ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
 	if (ret < 0)
@@ -4372,7 +4394,8 @@ int btrfs_write_inode(struct inode *inode, struct writeback_control *wbc)
 		return 0;
 
 	smp_mb();
-	nolock = (root->fs_info->closing && root == root->fs_info->tree_root);
+	if (root->fs_info->closing && is_free_space_inode(root, inode))
+		nolock = true;
 
 	if (wbc->sync_mode == WB_SYNC_ALL) {
 		if (nolock)
@@ -4417,8 +4440,9 @@ void btrfs_dirty_inode(struct inode *inode)
 		if (IS_ERR(trans)) {
 			if (printk_ratelimit()) {
 				printk(KERN_ERR "btrfs: fail to "
-				       "dirty  inode %lu error %ld\n",
-				       inode->i_ino, PTR_ERR(trans));
+				       "dirty  inode %llu error %ld\n",
+				       (unsigned long long)btrfs_ino(inode),
+				       PTR_ERR(trans));
 			}
 			return;
 		}
@@ -4428,8 +4452,9 @@ void btrfs_dirty_inode(struct inode *inode)
 		if (ret) {
 			if (printk_ratelimit()) {
 				printk(KERN_ERR "btrfs: fail to "
-				       "dirty  inode %lu error %d\n",
-				       inode->i_ino, ret);
+				       "dirty  inode %llu error %d\n",
+				       (unsigned long long)btrfs_ino(inode),
+				       ret);
 			}
 		}
 	}
@@ -4449,7 +4474,7 @@ static int btrfs_set_inode_index_count(struct inode *inode)
 	struct extent_buffer *leaf;
 	int ret;
 
-	key.objectid = inode->i_ino;
+	key.objectid = btrfs_ino(inode);
 	btrfs_set_key_type(&key, BTRFS_DIR_INDEX_KEY);
 	key.offset = (u64)-1;
 
@@ -4481,7 +4506,7 @@ static int btrfs_set_inode_index_count(struct inode *inode)
 	leaf = path->nodes[0];
 	btrfs_item_key_to_cpu(leaf, &found_key, path->slots[0]);
 
-	if (found_key.objectid != inode->i_ino ||
+	if (found_key.objectid != btrfs_ino(inode) ||
 	    btrfs_key_type(&found_key) != BTRFS_DIR_INDEX_KEY) {
 		BTRFS_I(inode)->index_cnt = 2;
 		goto out;
@@ -4540,6 +4565,12 @@ static struct inode *btrfs_new_inode(struct btrfs_trans_handle *trans,
 		return ERR_PTR(-ENOMEM);
 	}
 
+	/*
+	 * we have to initialize this early, so we can reclaim the inode
+	 * number if we fail afterwards in this function.
+	 */
+	inode->i_ino = objectid;
+
 	if (dir) {
 		trace_btrfs_inode_request(dir);
 
@@ -4585,7 +4616,6 @@ static struct inode *btrfs_new_inode(struct btrfs_trans_handle *trans,
 		goto fail;
 
 	inode_init_owner(inode, dir, mode);
-	inode->i_ino = objectid;
 	inode_set_bytes(inode, 0);
 	inode->i_mtime = inode->i_atime = inode->i_ctime = CURRENT_TIME;
 	inode_item = btrfs_item_ptr(path->nodes[0], path->slots[0],
@@ -4649,29 +4679,29 @@ int btrfs_add_link(struct btrfs_trans_handle *trans,
 	int ret = 0;
 	struct btrfs_key key;
 	struct btrfs_root *root = BTRFS_I(parent_inode)->root;
+	u64 ino = btrfs_ino(inode);
+	u64 parent_ino = btrfs_ino(parent_inode);
 
-	if (unlikely(inode->i_ino == BTRFS_FIRST_FREE_OBJECTID)) {
+	if (unlikely(ino == BTRFS_FIRST_FREE_OBJECTID)) {
 		memcpy(&key, &BTRFS_I(inode)->root->root_key, sizeof(key));
 	} else {
-		key.objectid = inode->i_ino;
+		key.objectid = ino;
 		btrfs_set_key_type(&key, BTRFS_INODE_ITEM_KEY);
 		key.offset = 0;
 	}
 
-	if (unlikely(inode->i_ino == BTRFS_FIRST_FREE_OBJECTID)) {
+	if (unlikely(ino == BTRFS_FIRST_FREE_OBJECTID)) {
 		ret = btrfs_add_root_ref(trans, root->fs_info->tree_root,
 					 key.objectid, root->root_key.objectid,
-					 parent_inode->i_ino,
-					 index, name, name_len);
+					 parent_ino, index, name, name_len);
 	} else if (add_backref) {
-		ret = btrfs_insert_inode_ref(trans, root,
-					     name, name_len, inode->i_ino,
-					     parent_inode->i_ino, index);
+		ret = btrfs_insert_inode_ref(trans, root, name, name_len, ino,
+					     parent_ino, index);
 	}
 
 	if (ret == 0) {
 		ret = btrfs_insert_dir_item(trans, root, name, name_len,
-					    parent_inode->i_ino, &key,
+					    parent_ino, &key,
 					    btrfs_inode_type(inode), index);
 		BUG_ON(ret);
 
@@ -4714,10 +4744,6 @@ static int btrfs_mknod(struct inode *dir, struct dentry *dentry,
 	if (!new_valid_dev(rdev))
 		return -EINVAL;
 
-	err = btrfs_find_free_objectid(NULL, root, dir->i_ino, &objectid);
-	if (err)
-		return err;
-
 	/*
 	 * 2 for inode item and ref
 	 * 2 for dir items
@@ -4729,8 +4755,12 @@ static int btrfs_mknod(struct inode *dir, struct dentry *dentry,
 
 	btrfs_set_trans_block_group(trans, dir);
 
+	err = btrfs_find_free_ino(root, &objectid);
+	if (err)
+		goto out_unlock;
+
 	inode = btrfs_new_inode(trans, root, dir, dentry->d_name.name,
-				dentry->d_name.len, dir->i_ino, objectid,
+				dentry->d_name.len, btrfs_ino(dir), objectid,
 				BTRFS_I(dir)->block_group, mode, &index);
 	if (IS_ERR(inode)) {
 		err = PTR_ERR(inode);
@@ -4777,9 +4807,6 @@ static int btrfs_create(struct inode *dir, struct dentry *dentry,
 	u64 objectid;
 	u64 index = 0;
 
-	err = btrfs_find_free_objectid(NULL, root, dir->i_ino, &objectid);
-	if (err)
-		return err;
 	/*
 	 * 2 for inode item and ref
 	 * 2 for dir items
@@ -4791,8 +4818,12 @@ static int btrfs_create(struct inode *dir, struct dentry *dentry,
 
 	btrfs_set_trans_block_group(trans, dir);
 
+	err = btrfs_find_free_ino(root, &objectid);
+	if (err)
+		goto out_unlock;
+
 	inode = btrfs_new_inode(trans, root, dir, dentry->d_name.name,
-				dentry->d_name.len, dir->i_ino, objectid,
+				dentry->d_name.len, btrfs_ino(dir), objectid,
 				BTRFS_I(dir)->block_group, mode, &index);
 	if (IS_ERR(inode)) {
 		err = PTR_ERR(inode);
@@ -4903,10 +4934,6 @@ static int btrfs_mkdir(struct inode *dir, struct dentry *dentry, int mode)
 	u64 index = 0;
 	unsigned long nr = 1;
 
-	err = btrfs_find_free_objectid(NULL, root, dir->i_ino, &objectid);
-	if (err)
-		return err;
-
 	/*
 	 * 2 items for inode and ref
 	 * 2 items for dir items
@@ -4917,8 +4944,12 @@ static int btrfs_mkdir(struct inode *dir, struct dentry *dentry, int mode)
 		return PTR_ERR(trans);
 	btrfs_set_trans_block_group(trans, dir);
 
+	err = btrfs_find_free_ino(root, &objectid);
+	if (err)
+		goto out_fail;
+
 	inode = btrfs_new_inode(trans, root, dir, dentry->d_name.name,
-				dentry->d_name.len, dir->i_ino, objectid,
+				dentry->d_name.len, btrfs_ino(dir), objectid,
 				BTRFS_I(dir)->block_group, S_IFDIR | mode,
 				&index);
 	if (IS_ERR(inode)) {
@@ -5041,7 +5072,7 @@ struct extent_map *btrfs_get_extent(struct inode *inode, struct page *page,
 	u64 bytenr;
 	u64 extent_start = 0;
 	u64 extent_end = 0;
-	u64 objectid = inode->i_ino;
+	u64 objectid = btrfs_ino(inode);
 	u32 found_type;
 	struct btrfs_path *path = NULL;
 	struct btrfs_root *root = BTRFS_I(inode)->root;
@@ -5549,7 +5580,7 @@ static noinline int can_nocow_odirect(struct btrfs_trans_handle *trans,
 	if (!path)
 		return -ENOMEM;
 
-	ret = btrfs_lookup_file_extent(trans, root, path, inode->i_ino,
+	ret = btrfs_lookup_file_extent(trans, root, path, btrfs_ino(inode),
 				       offset, 0);
 	if (ret < 0)
 		goto out;
@@ -5566,7 +5597,7 @@ static noinline int can_nocow_odirect(struct btrfs_trans_handle *trans,
 	ret = 0;
 	leaf = path->nodes[0];
 	btrfs_item_key_to_cpu(leaf, &key, slot);
-	if (key.objectid != inode->i_ino ||
+	if (key.objectid != btrfs_ino(inode) ||
 	    key.type != BTRFS_EXTENT_DATA_KEY) {
 		/* not our file or wrong item type, must cow */
 		goto out;
@@ -5600,7 +5631,7 @@ static noinline int can_nocow_odirect(struct btrfs_trans_handle *trans,
 	 * look for other files referencing this extent, if we
 	 * find any we must cow
 	 */
-	if (btrfs_cross_ref_exist(trans, root, inode->i_ino,
+	if (btrfs_cross_ref_exist(trans, root, btrfs_ino(inode),
 				  key.offset - backref_offset, disk_bytenr))
 		goto out;
 
@@ -5790,9 +5821,10 @@ static void btrfs_endio_direct_read(struct bio *bio, int err)
 
 			flush_dcache_page(bvec->bv_page);
 			if (csum != *private) {
-				printk(KERN_ERR "btrfs csum failed ino %lu off"
+				printk(KERN_ERR "btrfs csum failed ino %llu off"
 				      " %llu csum %u private %u\n",
-				      inode->i_ino, (unsigned long long)start,
+				      (unsigned long long)btrfs_ino(inode),
+				      (unsigned long long)start,
 				      csum, *private);
 				err = -EIO;
 			}
@@ -5939,9 +5971,9 @@ static void btrfs_end_dio_bio(struct bio *bio, int err)
 	struct btrfs_dio_private *dip = bio->bi_private;
 
 	if (err) {
-		printk(KERN_ERR "btrfs direct IO failed ino %lu rw %lu "
+		printk(KERN_ERR "btrfs direct IO failed ino %llu rw %lu "
 		      "sector %#Lx len %u err no %d\n",
-		      dip->inode->i_ino, bio->bi_rw,
+		      (unsigned long long)btrfs_ino(dip->inode), bio->bi_rw,
 		      (unsigned long long)bio->bi_sector, bio->bi_size, err);
 		dip->errors = 1;
 
@@ -6851,8 +6883,8 @@ void btrfs_destroy_inode(struct inode *inode)
 
 	spin_lock(&root->orphan_lock);
 	if (!list_empty(&BTRFS_I(inode)->i_orphan)) {
-		printk(KERN_INFO "BTRFS: inode %lu still on the orphan list\n",
-		       inode->i_ino);
+		printk(KERN_INFO "BTRFS: inode %llu still on the orphan list\n",
+		       (unsigned long long)btrfs_ino(inode));
 		list_del_init(&BTRFS_I(inode)->i_orphan);
 	}
 	spin_unlock(&root->orphan_lock);
@@ -6882,7 +6914,7 @@ int btrfs_drop_inode(struct inode *inode)
 	struct btrfs_root *root = BTRFS_I(inode)->root;
 
 	if (btrfs_root_refs(&root->root_item) == 0 &&
-	    root != root->fs_info->tree_root)
+	    !is_free_space_inode(root, inode))
 		return 1;
 	else
 		return generic_drop_inode(inode);
@@ -6991,16 +7023,17 @@ static int btrfs_rename(struct inode *old_dir, struct dentry *old_dentry,
 	u64 index = 0;
 	u64 root_objectid;
 	int ret;
+	u64 old_ino = btrfs_ino(old_inode);
 
-	if (new_dir->i_ino == BTRFS_EMPTY_SUBVOL_DIR_OBJECTID)
+	if (btrfs_ino(new_dir) == BTRFS_EMPTY_SUBVOL_DIR_OBJECTID)
 		return -EPERM;
 
 	/* we only allow rename subvolume link between subvolumes */
-	if (old_inode->i_ino != BTRFS_FIRST_FREE_OBJECTID && root != dest)
+	if (old_ino != BTRFS_FIRST_FREE_OBJECTID && root != dest)
 		return -EXDEV;
 
-	if (old_inode->i_ino == BTRFS_EMPTY_SUBVOL_DIR_OBJECTID ||
-	    (new_inode && new_inode->i_ino == BTRFS_FIRST_FREE_OBJECTID))
+	if (old_ino == BTRFS_EMPTY_SUBVOL_DIR_OBJECTID ||
+	    (new_inode && btrfs_ino(new_inode) == BTRFS_FIRST_FREE_OBJECTID))
 		return -ENOTEMPTY;
 
 	if (S_ISDIR(old_inode->i_mode) && new_inode &&
@@ -7016,7 +7049,7 @@ static int btrfs_rename(struct inode *old_dir, struct dentry *old_dentry,
 		filemap_flush(old_inode->i_mapping);
 
 	/* close the racy window with snapshot create/destroy ioctl */
-	if (old_inode->i_ino == BTRFS_FIRST_FREE_OBJECTID)
+	if (old_ino == BTRFS_FIRST_FREE_OBJECTID)
 		down_read(&root->fs_info->subvol_sem);
 	/*
 	 * We want to reserve the absolute worst case amount of items.  So if
@@ -7041,15 +7074,15 @@ static int btrfs_rename(struct inode *old_dir, struct dentry *old_dentry,
 	if (ret)
 		goto out_fail;
 
-	if (unlikely(old_inode->i_ino == BTRFS_FIRST_FREE_OBJECTID)) {
+	if (unlikely(old_ino == BTRFS_FIRST_FREE_OBJECTID)) {
 		/* force full log commit if subvolume involved. */
 		root->fs_info->last_trans_log_full_commit = trans->transid;
 	} else {
 		ret = btrfs_insert_inode_ref(trans, dest,
 					     new_dentry->d_name.name,
 					     new_dentry->d_name.len,
-					     old_inode->i_ino,
-					     new_dir->i_ino, index);
+					     old_ino,
+					     btrfs_ino(new_dir), index);
 		if (ret)
 			goto out_fail;
 		/*
@@ -7065,10 +7098,8 @@ static int btrfs_rename(struct inode *old_dir, struct dentry *old_dentry,
 	 * make sure the inode gets flushed if it is replacing
 	 * something.
 	 */
-	if (new_inode && new_inode->i_size &&
-	    old_inode && S_ISREG(old_inode->i_mode)) {
+	if (new_inode && new_inode->i_size && S_ISREG(old_inode->i_mode))
 		btrfs_add_ordered_operation(trans, root, old_inode);
-	}
 
 	old_dir->i_ctime = old_dir->i_mtime = ctime;
 	new_dir->i_ctime = new_dir->i_mtime = ctime;
@@ -7077,7 +7108,7 @@ static int btrfs_rename(struct inode *old_dir, struct dentry *old_dentry,
 	if (old_dentry->d_parent != new_dentry->d_parent)
 		btrfs_record_unlink_dir(trans, old_dir, old_inode, 1);
 
-	if (unlikely(old_inode->i_ino == BTRFS_FIRST_FREE_OBJECTID)) {
+	if (unlikely(old_ino == BTRFS_FIRST_FREE_OBJECTID)) {
 		root_objectid = BTRFS_I(old_inode)->root->root_key.objectid;
 		ret = btrfs_unlink_subvol(trans, root, old_dir, root_objectid,
 					old_dentry->d_name.name,
@@ -7094,7 +7125,7 @@ static int btrfs_rename(struct inode *old_dir, struct dentry *old_dentry,
 
 	if (new_inode) {
 		new_inode->i_ctime = CURRENT_TIME;
-		if (unlikely(new_inode->i_ino ==
+		if (unlikely(btrfs_ino(new_inode) ==
 			     BTRFS_EMPTY_SUBVOL_DIR_OBJECTID)) {
 			root_objectid = BTRFS_I(new_inode)->location.objectid;
 			ret = btrfs_unlink_subvol(trans, dest, new_dir,
@@ -7122,7 +7153,7 @@ static int btrfs_rename(struct inode *old_dir, struct dentry *old_dentry,
 			     new_dentry->d_name.len, 0, index);
 	BUG_ON(ret);
 
-	if (old_inode->i_ino != BTRFS_FIRST_FREE_OBJECTID) {
+	if (old_ino != BTRFS_FIRST_FREE_OBJECTID) {
 		struct dentry *parent = dget_parent(new_dentry);
 		btrfs_log_new_name(trans, old_inode, old_dir, parent);
 		dput(parent);
@@ -7131,7 +7162,7 @@ static int btrfs_rename(struct inode *old_dir, struct dentry *old_dentry,
 out_fail:
 	btrfs_end_transaction_throttle(trans, root);
 out_notrans:
-	if (old_inode->i_ino == BTRFS_FIRST_FREE_OBJECTID)
+	if (old_ino == BTRFS_FIRST_FREE_OBJECTID)
 		up_read(&root->fs_info->subvol_sem);
 
 	return ret;
@@ -7260,9 +7291,6 @@ static int btrfs_symlink(struct inode *dir, struct dentry *dentry,
 	if (name_len > BTRFS_MAX_INLINE_DATA_SIZE(root))
 		return -ENAMETOOLONG;
 
-	err = btrfs_find_free_objectid(NULL, root, dir->i_ino, &objectid);
-	if (err)
-		return err;
 	/*
 	 * 2 items for inode item and ref
 	 * 2 items for dir items
@@ -7274,8 +7302,12 @@ static int btrfs_symlink(struct inode *dir, struct dentry *dentry,
 
 	btrfs_set_trans_block_group(trans, dir);
 
+	err = btrfs_find_free_ino(root, &objectid);
+	if (err)
+		goto out_unlock;
+
 	inode = btrfs_new_inode(trans, root, dir, dentry->d_name.name,
-				dentry->d_name.len, dir->i_ino, objectid,
+				dentry->d_name.len, btrfs_ino(dir), objectid,
 				BTRFS_I(dir)->block_group, S_IFLNK|S_IRWXUGO,
 				&index);
 	if (IS_ERR(inode)) {
@@ -7307,7 +7339,7 @@ static int btrfs_symlink(struct inode *dir, struct dentry *dentry,
 
 	path = btrfs_alloc_path();
 	BUG_ON(!path);
-	key.objectid = inode->i_ino;
+	key.objectid = btrfs_ino(inode);
 	key.offset = 0;
 	btrfs_set_key_type(&key, BTRFS_EXTENT_DATA_KEY);
 	datasize = btrfs_file_extent_calc_inline_size(name_len);

fs/btrfs/ioctl.c

@@ -50,6 +50,7 @@
 #include "print-tree.h"
 #include "volumes.h"
 #include "locking.h"
+#include "inode-map.h"
 
 /* Mask out flags that are inappropriate for the given type of inode. */
 static inline __u32 btrfs_mask_flags(umode_t mode, __u32 flags)
@@ -329,8 +330,7 @@ static noinline int create_subvol(struct btrfs_root *root,
 	u64 new_dirid = BTRFS_FIRST_FREE_OBJECTID;
 	u64 index = 0;
 
-	ret = btrfs_find_free_objectid(NULL, root->fs_info->tree_root,
-				       0, &objectid);
+	ret = btrfs_find_free_objectid(root->fs_info->tree_root, &objectid);
 	if (ret) {
 		dput(parent);
 		return ret;
@@ -422,7 +422,7 @@ static noinline int create_subvol(struct btrfs_root *root,
 	BUG_ON(ret);
 
 	ret = btrfs_insert_dir_item(trans, root,
-				    name, namelen, dir->i_ino, &key,
+				    name, namelen, btrfs_ino(dir), &key,
 				    BTRFS_FT_DIR, index);
 	if (ret)
 		goto fail;
@@ -433,7 +433,7 @@ static noinline int create_subvol(struct btrfs_root *root,
 
 	ret = btrfs_add_root_ref(trans, root->fs_info->tree_root,
 				 objectid, root->root_key.objectid,
-				 dir->i_ino, index, name, namelen);
+				 btrfs_ino(dir), index, name, namelen);
 
 	BUG_ON(ret);
 
@@ -1129,7 +1129,7 @@ static noinline int btrfs_ioctl_subvol_getflags(struct file *file,
 	int ret = 0;
 	u64 flags = 0;
 
-	if (inode->i_ino != BTRFS_FIRST_FREE_OBJECTID)
+	if (btrfs_ino(inode) != BTRFS_FIRST_FREE_OBJECTID)
 		return -EINVAL;
 
 	down_read(&root->fs_info->subvol_sem);
@@ -1156,7 +1156,7 @@ static noinline int btrfs_ioctl_subvol_setflags(struct file *file,
 	if (root->fs_info->sb->s_flags & MS_RDONLY)
 		return -EROFS;
 
-	if (inode->i_ino != BTRFS_FIRST_FREE_OBJECTID)
+	if (btrfs_ino(inode) != BTRFS_FIRST_FREE_OBJECTID)
 		return -EINVAL;
 
 	if (copy_from_user(&flags, arg, sizeof(flags)))
@@ -1639,7 +1639,7 @@ static noinline int btrfs_ioctl_snap_destroy(struct file *file,
 			goto out_dput;
 	}
 
-	if (inode->i_ino != BTRFS_FIRST_FREE_OBJECTID) {
+	if (btrfs_ino(inode) != BTRFS_FIRST_FREE_OBJECTID) {
 		err = -EINVAL;
 		goto out_dput;
 	}
@@ -1925,7 +1925,7 @@ static noinline long btrfs_ioctl_clone(struct file *file, unsigned long srcfd,
 	}
 
 	/* clone data */
-	key.objectid = src->i_ino;
+	key.objectid = btrfs_ino(src);
 	key.type = BTRFS_EXTENT_DATA_KEY;
 	key.offset = 0;
 
@@ -1952,7 +1952,7 @@ static noinline long btrfs_ioctl_clone(struct file *file, unsigned long srcfd,
 
 		btrfs_item_key_to_cpu(leaf, &key, slot);
 		if (btrfs_key_type(&key) > BTRFS_EXTENT_DATA_KEY ||
-		    key.objectid != src->i_ino)
+		    key.objectid != btrfs_ino(src))
 			break;
 
 		if (btrfs_key_type(&key) == BTRFS_EXTENT_DATA_KEY) {
@@ -1995,7 +1995,7 @@ static noinline long btrfs_ioctl_clone(struct file *file, unsigned long srcfd,
 				goto next;
 
 			memcpy(&new_key, &key, sizeof(new_key));
-			new_key.objectid = inode->i_ino;
+			new_key.objectid = btrfs_ino(inode);
 			if (off <= key.offset)
 				new_key.offset = key.offset + destoff - off;
 			else
@@ -2049,7 +2049,7 @@ static noinline long btrfs_ioctl_clone(struct file *file, unsigned long srcfd,
 					ret = btrfs_inc_extent_ref(trans, root,
 							disko, diskl, 0,
 							root->root_key.objectid,
-							inode->i_ino,
+							btrfs_ino(inode),
 							new_key.offset - datao);
 					BUG_ON(ret);
 				}

fs/btrfs/relocation.c

@@ -30,6 +30,7 @@
 #include "btrfs_inode.h"
 #include "async-thread.h"
 #include "free-space-cache.h"
+#include "inode-map.h"
 
 /*
  * backref_node, mapping_node and tree_block start with this
@@ -1409,9 +1410,9 @@ again:
 		prev = node;
 		entry = rb_entry(node, struct btrfs_inode, rb_node);
 
-		if (objectid < entry->vfs_inode.i_ino)
+		if (objectid < btrfs_ino(&entry->vfs_inode))
 			node = node->rb_left;
-		else if (objectid > entry->vfs_inode.i_ino)
+		else if (objectid > btrfs_ino(&entry->vfs_inode))
 			node = node->rb_right;
 		else
 			break;
@@ -1419,7 +1420,7 @@ again:
 	if (!node) {
 		while (prev) {
 			entry = rb_entry(prev, struct btrfs_inode, rb_node);
-			if (objectid <= entry->vfs_inode.i_ino) {
+			if (objectid <= btrfs_ino(&entry->vfs_inode)) {
 				node = prev;
 				break;
 			}
@@ -1434,7 +1435,7 @@ again:
 			return inode;
 		}
 
-		objectid = entry->vfs_inode.i_ino + 1;
+		objectid = btrfs_ino(&entry->vfs_inode) + 1;
 		if (cond_resched_lock(&root->inode_lock))
 			goto again;
 
@@ -1470,7 +1471,7 @@ static int get_new_location(struct inode *reloc_inode, u64 *new_bytenr,
 		return -ENOMEM;
 
 	bytenr -= BTRFS_I(reloc_inode)->index_cnt;
-	ret = btrfs_lookup_file_extent(NULL, root, path, reloc_inode->i_ino,
+	ret = btrfs_lookup_file_extent(NULL, root, path, btrfs_ino(reloc_inode),
 				       bytenr, 0);
 	if (ret < 0)
 		goto out;
@@ -1558,11 +1559,11 @@ int replace_file_extents(struct btrfs_trans_handle *trans,
 			if (first) {
 				inode = find_next_inode(root, key.objectid);
 				first = 0;
-			} else if (inode && inode->i_ino < key.objectid) {
+			} else if (inode && btrfs_ino(inode) < key.objectid) {
 				btrfs_add_delayed_iput(inode);
 				inode = find_next_inode(root, key.objectid);
 			}
-			if (inode && inode->i_ino == key.objectid) {
+			if (inode && btrfs_ino(inode) == key.objectid) {
 				end = key.offset +
 				      btrfs_file_extent_num_bytes(leaf, fi);
 				WARN_ON(!IS_ALIGNED(key.offset,
@@ -1893,6 +1894,7 @@ static int invalidate_extent_cache(struct btrfs_root *root,
 	struct inode *inode = NULL;
 	u64 objectid;
 	u64 start, end;
+	u64 ino;
 
 	objectid = min_key->objectid;
 	while (1) {
@@ -1905,17 +1907,18 @@ static int invalidate_extent_cache(struct btrfs_root *root,
 		inode = find_next_inode(root, objectid);
 		if (!inode)
 			break;
+		ino = btrfs_ino(inode);
 
-		if (inode->i_ino > max_key->objectid) {
+		if (ino > max_key->objectid) {
 			iput(inode);
 			break;
 		}
 
-		objectid = inode->i_ino + 1;
+		objectid = ino + 1;
 		if (!S_ISREG(inode->i_mode))
 			continue;
 
-		if (unlikely(min_key->objectid == inode->i_ino)) {
+		if (unlikely(min_key->objectid == ino)) {
 			if (min_key->type > BTRFS_EXTENT_DATA_KEY)
 				continue;
 			if (min_key->type < BTRFS_EXTENT_DATA_KEY)
@@ -1928,7 +1931,7 @@ static int invalidate_extent_cache(struct btrfs_root *root,
 			start = 0;
 		}
 
-		if (unlikely(max_key->objectid == inode->i_ino)) {
+		if (unlikely(max_key->objectid == ino)) {
 			if (max_key->type < BTRFS_EXTENT_DATA_KEY)
 				continue;
 			if (max_key->type > BTRFS_EXTENT_DATA_KEY) {
@@ -3897,7 +3900,7 @@ struct inode *create_reloc_inode(struct btrfs_fs_info *fs_info,
 	if (IS_ERR(trans))
 		return ERR_CAST(trans);
 
-	err = btrfs_find_free_objectid(trans, root, objectid, &objectid);
+	err = btrfs_find_free_objectid(root, &objectid);
 	if (err)
 		goto out;
 

fs/btrfs/transaction.c

@@ -27,6 +27,7 @@
 #include "transaction.h"
 #include "locking.h"
 #include "tree-log.h"
+#include "inode-map.h"
 
 #define BTRFS_ROOT_TRANS_TAG 0
 
@@ -760,8 +761,14 @@ static noinline int commit_fs_roots(struct btrfs_trans_handle *trans,
 			btrfs_update_reloc_root(trans, root);
 			btrfs_orphan_commit_root(trans, root);
 
+			btrfs_save_ino_cache(root, trans);
+
 			if (root->commit_root != root->node) {
+				mutex_lock(&root->fs_commit_mutex);
 				switch_commit_root(root);
+				btrfs_unpin_free_ino(root);
+				mutex_unlock(&root->fs_commit_mutex);
+
 				btrfs_set_root_node(&root->root_item,
 						    root->node);
 			}
@@ -930,7 +937,7 @@ static noinline int create_pending_snapshot(struct btrfs_trans_handle *trans,
 		goto fail;
 	}
 
-	ret = btrfs_find_free_objectid(trans, tree_root, 0, &objectid);
+	ret = btrfs_find_free_objectid(tree_root, &objectid);
 	if (ret) {
 		pending->error = ret;
 		goto fail;
@@ -967,7 +974,7 @@ static noinline int create_pending_snapshot(struct btrfs_trans_handle *trans,
 	BUG_ON(ret);
 	ret = btrfs_insert_dir_item(trans, parent_root,
 				dentry->d_name.name, dentry->d_name.len,
-				parent_inode->i_ino, &key,
+				btrfs_ino(parent_inode), &key,
 				BTRFS_FT_DIR, index);
 	BUG_ON(ret);
 
@@ -1009,7 +1016,7 @@ static noinline int create_pending_snapshot(struct btrfs_trans_handle *trans,
 	 */
 	ret = btrfs_add_root_ref(trans, tree_root, objectid,
 				 parent_root->root_key.objectid,
-				 parent_inode->i_ino, index,
+				 btrfs_ino(parent_inode), index,
 				 dentry->d_name.name, dentry->d_name.len);
 	BUG_ON(ret);
 	dput(parent);

fs/btrfs/tree-log.c

@@ -519,7 +519,7 @@ static noinline int replay_one_extent(struct btrfs_trans_handle *trans,
 	 * file.  This must be done before the btrfs_drop_extents run
 	 * so we don't try to drop this extent.
 	 */
-	ret = btrfs_lookup_file_extent(trans, root, path, inode->i_ino,
+	ret = btrfs_lookup_file_extent(trans, root, path, btrfs_ino(inode),
 				       start, 0);
 
 	if (ret == 0 &&
@@ -832,7 +832,7 @@ again:
 	read_extent_buffer(eb, name, (unsigned long)(ref + 1), namelen);
 
 	/* if we already have a perfect match, we're done */
-	if (inode_in_dir(root, path, dir->i_ino, inode->i_ino,
+	if (inode_in_dir(root, path, btrfs_ino(dir), btrfs_ino(inode),
 			 btrfs_inode_ref_index(eb, ref),
 			 name, namelen)) {
 		goto out;
@@ -960,8 +960,9 @@ static noinline int fixup_inode_link_count(struct btrfs_trans_handle *trans,
 	unsigned long ptr;
 	unsigned long ptr_end;
 	int name_len;
+	u64 ino = btrfs_ino(inode);
 
-	key.objectid = inode->i_ino;
+	key.objectid = ino;
 	key.type = BTRFS_INODE_REF_KEY;
 	key.offset = (u64)-1;
 
@@ -980,7 +981,7 @@ static noinline int fixup_inode_link_count(struct btrfs_trans_handle *trans,
 		}
 		btrfs_item_key_to_cpu(path->nodes[0], &key,
 				      path->slots[0]);
-		if (key.objectid != inode->i_ino ||
+		if (key.objectid != ino ||
 		    key.type != BTRFS_INODE_REF_KEY)
 			break;
 		ptr = btrfs_item_ptr_offset(path->nodes[0], path->slots[0]);
@@ -1011,10 +1012,10 @@ static noinline int fixup_inode_link_count(struct btrfs_trans_handle *trans,
 	if (inode->i_nlink == 0) {
 		if (S_ISDIR(inode->i_mode)) {
 			ret = replay_dir_deletes(trans, root, NULL, path,
-						 inode->i_ino, 1);
+						 ino, 1);
 			BUG_ON(ret);
 		}
-		ret = insert_orphan_item(trans, root, inode->i_ino);
+		ret = insert_orphan_item(trans, root, ino);
 		BUG_ON(ret);
 	}
 	btrfs_free_path(path);
@@ -2197,6 +2198,7 @@ int btrfs_del_dir_entries_in_log(struct btrfs_trans_handle *trans,
 	int ret;
 	int err = 0;
 	int bytes_del = 0;
+	u64 dir_ino = btrfs_ino(dir);
 
 	if (BTRFS_I(dir)->logged_trans < trans->transid)
 		return 0;
@@ -2214,7 +2216,7 @@ int btrfs_del_dir_entries_in_log(struct btrfs_trans_handle *trans,
 		goto out_unlock;
 	}
 
-	di = btrfs_lookup_dir_item(trans, log, path, dir->i_ino,
+	di = btrfs_lookup_dir_item(trans, log, path, dir_ino,
 				   name, name_len, -1);
 	if (IS_ERR(di)) {
 		err = PTR_ERR(di);
@@ -2226,7 +2228,7 @@ int btrfs_del_dir_entries_in_log(struct btrfs_trans_handle *trans,
 		BUG_ON(ret);
 	}
 	btrfs_release_path(log, path);
-	di = btrfs_lookup_dir_index_item(trans, log, path, dir->i_ino,
+	di = btrfs_lookup_dir_index_item(trans, log, path, dir_ino,
 					 index, name, name_len, -1);
 	if (IS_ERR(di)) {
 		err = PTR_ERR(di);
@@ -2244,7 +2246,7 @@ int btrfs_del_dir_entries_in_log(struct btrfs_trans_handle *trans,
 	if (bytes_del) {
 		struct btrfs_key key;
 
-		key.objectid = dir->i_ino;
+		key.objectid = dir_ino;
 		key.offset = 0;
 		key.type = BTRFS_INODE_ITEM_KEY;
 		btrfs_release_path(log, path);
@@ -2303,7 +2305,7 @@ int btrfs_del_inode_ref_in_log(struct btrfs_trans_handle *trans,
 	log = root->log_root;
 	mutex_lock(&BTRFS_I(inode)->log_mutex);
 
-	ret = btrfs_del_inode_ref(trans, log, name, name_len, inode->i_ino,
+	ret = btrfs_del_inode_ref(trans, log, name, name_len, btrfs_ino(inode),
 				  dirid, &index);
 	mutex_unlock(&BTRFS_I(inode)->log_mutex);
 	if (ret == -ENOSPC) {
@@ -2369,13 +2371,14 @@ static noinline int log_dir_items(struct btrfs_trans_handle *trans,
 	int nritems;
 	u64 first_offset = min_offset;
 	u64 last_offset = (u64)-1;
+	u64 ino = btrfs_ino(inode);
 
 	log = root->log_root;
-	max_key.objectid = inode->i_ino;
+	max_key.objectid = ino;
 	max_key.offset = (u64)-1;
 	max_key.type = key_type;
 
-	min_key.objectid = inode->i_ino;
+	min_key.objectid = ino;
 	min_key.type = key_type;
 	min_key.offset = min_offset;
 
@@ -2388,9 +2391,8 @@ static noinline int log_dir_items(struct btrfs_trans_handle *trans,
 	 * we didn't find anything from this transaction, see if there
 	 * is anything at all
 	 */
-	if (ret != 0 || min_key.objectid != inode->i_ino ||
-	    min_key.type != key_type) {
-		min_key.objectid = inode->i_ino;
+	if (ret != 0 || min_key.objectid != ino || min_key.type != key_type) {
+		min_key.objectid = ino;
 		min_key.type = key_type;
 		min_key.offset = (u64)-1;
 		btrfs_release_path(root, path);
@@ -2399,7 +2401,7 @@ static noinline int log_dir_items(struct btrfs_trans_handle *trans,
 			btrfs_release_path(root, path);
 			return ret;
 		}
-		ret = btrfs_previous_item(root, path, inode->i_ino, key_type);
+		ret = btrfs_previous_item(root, path, ino, key_type);
 
 		/* if ret == 0 there are items for this type,
 		 * create a range to tell us the last key of this type.
@@ -2417,7 +2419,7 @@ static noinline int log_dir_items(struct btrfs_trans_handle *trans,
 	}
 
 	/* go backward to find any previous key */
-	ret = btrfs_previous_item(root, path, inode->i_ino, key_type);
+	ret = btrfs_previous_item(root, path, ino, key_type);
 	if (ret == 0) {
 		struct btrfs_key tmp;
 		btrfs_item_key_to_cpu(path->nodes[0], &tmp, path->slots[0]);
@@ -2452,8 +2454,7 @@ static noinline int log_dir_items(struct btrfs_trans_handle *trans,
 		for (i = path->slots[0]; i < nritems; i++) {
 			btrfs_item_key_to_cpu(src, &min_key, i);
 
-			if (min_key.objectid != inode->i_ino ||
-			    min_key.type != key_type)
+			if (min_key.objectid != ino || min_key.type != key_type)
 				goto done;
 			ret = overwrite_item(trans, log, dst_path, src, i,
 					     &min_key);
@@ -2474,7 +2475,7 @@ static noinline int log_dir_items(struct btrfs_trans_handle *trans,
 			goto done;
 		}
 		btrfs_item_key_to_cpu(path->nodes[0], &tmp, path->slots[0]);
-		if (tmp.objectid != inode->i_ino || tmp.type != key_type) {
+		if (tmp.objectid != ino || tmp.type != key_type) {
 			last_offset = (u64)-1;
 			goto done;
 		}
@@ -2500,8 +2501,7 @@ done:
 		 * is valid
 		 */
 		ret = insert_dir_log_key(trans, log, path, key_type,
-					 inode->i_ino, first_offset,
-					 last_offset);
+					 ino, first_offset, last_offset);
 		if (ret)
 			err = ret;
 	}
@@ -2745,6 +2745,7 @@ static int btrfs_log_inode(struct btrfs_trans_handle *trans,
 	int nritems;
 	int ins_start_slot = 0;
 	int ins_nr;
+	u64 ino = btrfs_ino(inode);
 
 	log = root->log_root;
 
@@ -2757,11 +2758,11 @@ static int btrfs_log_inode(struct btrfs_trans_handle *trans,
 		return -ENOMEM;
 	}
 
-	min_key.objectid = inode->i_ino;
+	min_key.objectid = ino;
 	min_key.type = BTRFS_INODE_ITEM_KEY;
 	min_key.offset = 0;
 
-	max_key.objectid = inode->i_ino;
+	max_key.objectid = ino;
 
 	/* today the code can only do partial logging of directories */
 	if (!S_ISDIR(inode->i_mode))
@@ -2784,8 +2785,7 @@ static int btrfs_log_inode(struct btrfs_trans_handle *trans,
 
 		if (inode_only == LOG_INODE_EXISTS)
 			max_key_type = BTRFS_XATTR_ITEM_KEY;
-		ret = drop_objectid_items(trans, log, path,
-					  inode->i_ino, max_key_type);
+		ret = drop_objectid_items(trans, log, path, ino, max_key_type);
 	} else {
 		ret = btrfs_truncate_inode_items(trans, log, inode, 0, 0);
 	}
@@ -2803,7 +2803,7 @@ static int btrfs_log_inode(struct btrfs_trans_handle *trans,
 			break;
 again:
 		/* note, ins_nr might be > 0 here, cleanup outside the loop */
-		if (min_key.objectid != inode->i_ino)
+		if (min_key.objectid != ino)
 			break;
 		if (min_key.type > max_key.type)
 			break;

fs/btrfs/xattr.c

@@ -44,7 +44,7 @@ ssize_t __btrfs_getxattr(struct inode *inode, const char *name,
 		return -ENOMEM;
 
 	/* lookup the xattr by name */
-	di = btrfs_lookup_xattr(NULL, root, path, inode->i_ino, name,
+	di = btrfs_lookup_xattr(NULL, root, path, btrfs_ino(inode), name,
 				strlen(name), 0);
 	if (!di) {
 		ret = -ENODATA;
@@ -103,7 +103,7 @@ static int do_setxattr(struct btrfs_trans_handle *trans,
 		return -ENOMEM;
 
 	/* first lets see if we already have this xattr */
-	di = btrfs_lookup_xattr(trans, root, path, inode->i_ino, name,
+	di = btrfs_lookup_xattr(trans, root, path, btrfs_ino(inode), name,
 				strlen(name), -1);
 	if (IS_ERR(di)) {
 		ret = PTR_ERR(di);
@@ -136,7 +136,7 @@ static int do_setxattr(struct btrfs_trans_handle *trans,
 	}
 
 	/* ok we have to create a completely new xattr */
-	ret = btrfs_insert_xattr_item(trans, root, path, inode->i_ino,
+	ret = btrfs_insert_xattr_item(trans, root, path, btrfs_ino(inode),
 				      name, name_len, value, size);
 	BUG_ON(ret);
 out:
@@ -190,7 +190,7 @@ ssize_t btrfs_listxattr(struct dentry *dentry, char *buffer, size_t size)
 	 * NOTE: we set key.offset = 0; because we want to start with the
 	 * first xattr that we find and walk forward
 	 */
-	key.objectid = inode->i_ino;
+	key.objectid = btrfs_ino(inode);
 	btrfs_set_key_type(&key, BTRFS_XATTR_ITEM_KEY);
 	key.offset = 0;