Merge commit 'v2.6.29-rc7' into perfcounters/core

Conflicts:
	arch/x86/mm/iomap_32.c

23 files changed, 436 insertions, 123 deletions

diff --git a/fs/Makefile b/fs/Makefile
index 38bc735c67ad..dc20db348679 100644
--- a/fs/Makefile
+++ b/fs/Makefile
@@ -69,10 +69,12 @@ obj-$(CONFIG_DLM)		+= dlm/
 # Do not add any filesystems before this line
 obj-$(CONFIG_REISERFS_FS)	+= reiserfs/
 obj-$(CONFIG_EXT3_FS)		+= ext3/ # Before ext2 so root fs can be ext3
-obj-$(CONFIG_EXT4_FS)		+= ext4/ # Before ext2 so root fs can be ext4
+obj-$(CONFIG_EXT2_FS)		+= ext2/
+# We place ext4 after ext2 so plain ext2 root fs's are mounted using ext2
+# unless explicitly requested by rootfstype
+obj-$(CONFIG_EXT4_FS)		+= ext4/
 obj-$(CONFIG_JBD)		+= jbd/
 obj-$(CONFIG_JBD2)		+= jbd2/
-obj-$(CONFIG_EXT2_FS)		+= ext2/
 obj-$(CONFIG_CRAMFS)		+= cramfs/
 obj-$(CONFIG_SQUASHFS)		+= squashfs/
 obj-y				+= ramfs/
diff --git a/fs/bio.c b/fs/bio.c
index 72ab251cdb9c..124b95c4d582 100644
--- a/fs/bio.c
+++ b/fs/bio.c
@@ -302,7 +302,7 @@ void bio_init(struct bio *bio)
 struct bio *bio_alloc_bioset(gfp_t gfp_mask, int nr_iovecs, struct bio_set *bs)
 {
 	struct bio *bio = NULL;
-	void *p;
+	void *uninitialized_var(p);
 
 	if (bs) {
 		p = mempool_alloc(bs->bio_pool, gfp_mask);
diff --git a/fs/btrfs/btrfs_inode.h b/fs/btrfs/btrfs_inode.h
index a8c9693b75ac..72677ce2b74f 100644
--- a/fs/btrfs/btrfs_inode.h
+++ b/fs/btrfs/btrfs_inode.h
@@ -66,6 +66,9 @@ struct btrfs_inode {
 	 */
 	struct list_head delalloc_inodes;
 
+	/* the space_info for where this inode's data allocations are done */
+	struct btrfs_space_info *space_info;
+
 	/* full 64 bit generation number, struct vfs_inode doesn't have a big
 	 * enough field for this.
 	 */
@@ -94,6 +97,11 @@ struct btrfs_inode {
 	 */
 	u64 delalloc_bytes;
 
+	/* total number of bytes that may be used for this inode for
+	 * delalloc
+	 */
+	u64 reserved_bytes;
+
 	/*
 	 * the size of the file stored in the metadata on disk.  data=ordered
 	 * means the in-memory i_size might be larger than the size on disk
diff --git a/fs/btrfs/ctree.h b/fs/btrfs/ctree.h
index 766b31ae3186..82491ba8fa40 100644
--- a/fs/btrfs/ctree.h
+++ b/fs/btrfs/ctree.h
@@ -596,13 +596,27 @@ struct btrfs_block_group_item {
 
 struct btrfs_space_info {
 	u64 flags;
-	u64 total_bytes;
-	u64 bytes_used;
-	u64 bytes_pinned;
-	u64 bytes_reserved;
-	u64 bytes_readonly;
-	int full;
-	int force_alloc;
+
+	u64 total_bytes;	/* total bytes in the space */
+	u64 bytes_used;		/* total bytes used on disk */
+	u64 bytes_pinned;	/* total bytes pinned, will be freed when the
+				   transaction finishes */
+	u64 bytes_reserved;	/* total bytes the allocator has reserved for
+				   current allocations */
+	u64 bytes_readonly;	/* total bytes that are read only */
+
+	/* delalloc accounting */
+	u64 bytes_delalloc;	/* number of bytes reserved for allocation,
+				   this space is not necessarily reserved yet
+				   by the allocator */
+	u64 bytes_may_use;	/* number of bytes that may be used for
+				   delalloc */
+
+	int full;		/* indicates that we cannot allocate any more
+				   chunks for this space */
+	int force_alloc;	/* set if we need to force a chunk alloc for
+				   this space */
+
 	struct list_head list;
 
 	/* for block groups in our same type */
@@ -1782,6 +1796,16 @@ int btrfs_add_dead_reloc_root(struct btrfs_root *root);
 int btrfs_cleanup_reloc_trees(struct btrfs_root *root);
 int btrfs_reloc_clone_csums(struct inode *inode, u64 file_pos, u64 len);
 u64 btrfs_reduce_alloc_profile(struct btrfs_root *root, u64 flags);
+void btrfs_set_inode_space_info(struct btrfs_root *root, struct inode *ionde);
+int btrfs_check_metadata_free_space(struct btrfs_root *root);
+int btrfs_check_data_free_space(struct btrfs_root *root, struct inode *inode,
+				u64 bytes);
+void btrfs_free_reserved_data_space(struct btrfs_root *root,
+				    struct inode *inode, u64 bytes);
+void btrfs_delalloc_reserve_space(struct btrfs_root *root, struct inode *inode,
+				 u64 bytes);
+void btrfs_delalloc_free_space(struct btrfs_root *root, struct inode *inode,
+			      u64 bytes);
 /* ctree.c */
 int btrfs_previous_item(struct btrfs_root *root,
 			struct btrfs_path *path, u64 min_objectid,
@@ -2027,8 +2051,6 @@ int btrfs_merge_bio_hook(struct page *page, unsigned long offset,
 unsigned long btrfs_force_ra(struct address_space *mapping,
 			      struct file_ra_state *ra, struct file *file,
 			      pgoff_t offset, pgoff_t last_index);
-int btrfs_check_free_space(struct btrfs_root *root, u64 num_required,
-			   int for_del);
 int btrfs_page_mkwrite(struct vm_area_struct *vma, struct page *page);
 int btrfs_readpage(struct file *file, struct page *page);
 void btrfs_delete_inode(struct inode *inode);
diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c
index 0a5d796c9f7e..6b5966aacf44 100644
--- a/fs/btrfs/extent-tree.c
+++ b/fs/btrfs/extent-tree.c
@@ -60,6 +60,10 @@ static int update_block_group(struct btrfs_trans_handle *trans,
 			      u64 bytenr, u64 num_bytes, int alloc,
 			      int mark_free);
 
+static int do_chunk_alloc(struct btrfs_trans_handle *trans,
+			  struct btrfs_root *extent_root, u64 alloc_bytes,
+			  u64 flags, int force);
+
 static int block_group_bits(struct btrfs_block_group_cache *cache, u64 bits)
 {
 	return (cache->flags & bits) == bits;
@@ -1909,6 +1913,7 @@ static int update_space_info(struct btrfs_fs_info *info, u64 flags,
 	found->bytes_pinned = 0;
 	found->bytes_reserved = 0;
 	found->bytes_readonly = 0;
+	found->bytes_delalloc = 0;
 	found->full = 0;
 	found->force_alloc = 0;
 	*space_info = found;
@@ -1972,6 +1977,233 @@ u64 btrfs_reduce_alloc_profile(struct btrfs_root *root, u64 flags)
 	return flags;
 }
 
+static u64 btrfs_get_alloc_profile(struct btrfs_root *root, u64 data)
+{
+	struct btrfs_fs_info *info = root->fs_info;
+	u64 alloc_profile;
+
+	if (data) {
+		alloc_profile = info->avail_data_alloc_bits &
+			info->data_alloc_profile;
+		data = BTRFS_BLOCK_GROUP_DATA | alloc_profile;
+	} else if (root == root->fs_info->chunk_root) {
+		alloc_profile = info->avail_system_alloc_bits &
+			info->system_alloc_profile;
+		data = BTRFS_BLOCK_GROUP_SYSTEM | alloc_profile;
+	} else {
+		alloc_profile = info->avail_metadata_alloc_bits &
+			info->metadata_alloc_profile;
+		data = BTRFS_BLOCK_GROUP_METADATA | alloc_profile;
+	}
+
+	return btrfs_reduce_alloc_profile(root, data);
+}
+
+void btrfs_set_inode_space_info(struct btrfs_root *root, struct inode *inode)
+{
+	u64 alloc_target;
+
+	alloc_target = btrfs_get_alloc_profile(root, 1);
+	BTRFS_I(inode)->space_info = __find_space_info(root->fs_info,
+						       alloc_target);
+}
+
+/*
+ * for now this just makes sure we have at least 5% of our metadata space free
+ * for use.
+ */
+int btrfs_check_metadata_free_space(struct btrfs_root *root)
+{
+	struct btrfs_fs_info *info = root->fs_info;
+	struct btrfs_space_info *meta_sinfo;
+	u64 alloc_target, thresh;
+	int committed = 0, ret;
+
+	/* get the space info for where the metadata will live */
+	alloc_target = btrfs_get_alloc_profile(root, 0);
+	meta_sinfo = __find_space_info(info, alloc_target);
+
+again:
+	spin_lock(&meta_sinfo->lock);
+	if (!meta_sinfo->full)
+		thresh = meta_sinfo->total_bytes * 80;
+	else
+		thresh = meta_sinfo->total_bytes * 95;
+
+	do_div(thresh, 100);
+
+	if (meta_sinfo->bytes_used + meta_sinfo->bytes_reserved +
+	    meta_sinfo->bytes_pinned + meta_sinfo->bytes_readonly > thresh) {
+		struct btrfs_trans_handle *trans;
+		if (!meta_sinfo->full) {
+			meta_sinfo->force_alloc = 1;
+			spin_unlock(&meta_sinfo->lock);
+
+			trans = btrfs_start_transaction(root, 1);
+			if (!trans)
+				return -ENOMEM;
+
+			ret = do_chunk_alloc(trans, root->fs_info->extent_root,
+					     2 * 1024 * 1024, alloc_target, 0);
+			btrfs_end_transaction(trans, root);
+			goto again;
+		}
+		spin_unlock(&meta_sinfo->lock);
+
+		if (!committed) {
+			committed = 1;
+			trans = btrfs_join_transaction(root, 1);
+			if (!trans)
+				return -ENOMEM;
+			ret = btrfs_commit_transaction(trans, root);
+			if (ret)
+				return ret;
+			goto again;
+		}
+		return -ENOSPC;
+	}
+	spin_unlock(&meta_sinfo->lock);
+
+	return 0;
+}
+
+/*
+ * This will check the space that the inode allocates from to make sure we have
+ * enough space for bytes.
+ */
+int btrfs_check_data_free_space(struct btrfs_root *root, struct inode *inode,
+				u64 bytes)
+{
+	struct btrfs_space_info *data_sinfo;
+	int ret = 0, committed = 0;
+
+	/* make sure bytes are sectorsize aligned */
+	bytes = (bytes + root->sectorsize - 1) & ~((u64)root->sectorsize - 1);
+
+	data_sinfo = BTRFS_I(inode)->space_info;
+again:
+	/* make sure we have enough space to handle the data first */
+	spin_lock(&data_sinfo->lock);
+	if (data_sinfo->total_bytes - data_sinfo->bytes_used -
+	    data_sinfo->bytes_delalloc - data_sinfo->bytes_reserved -
+	    data_sinfo->bytes_pinned - data_sinfo->bytes_readonly -
+	    data_sinfo->bytes_may_use < bytes) {
+		struct btrfs_trans_handle *trans;
+
+		/*
+		 * if we don't have enough free bytes in this space then we need
+		 * to alloc a new chunk.
+		 */
+		if (!data_sinfo->full) {
+			u64 alloc_target;
+
+			data_sinfo->force_alloc = 1;
+			spin_unlock(&data_sinfo->lock);
+
+			alloc_target = btrfs_get_alloc_profile(root, 1);
+			trans = btrfs_start_transaction(root, 1);
+			if (!trans)
+				return -ENOMEM;
+
+			ret = do_chunk_alloc(trans, root->fs_info->extent_root,
+					     bytes + 2 * 1024 * 1024,
+					     alloc_target, 0);
+			btrfs_end_transaction(trans, root);
+			if (ret)
+				return ret;
+			goto again;
+		}
+		spin_unlock(&data_sinfo->lock);
+
+		/* commit the current transaction and try again */
+		if (!committed) {
+			committed = 1;
+			trans = btrfs_join_transaction(root, 1);
+			if (!trans)
+				return -ENOMEM;
+			ret = btrfs_commit_transaction(trans, root);
+			if (ret)
+				return ret;
+			goto again;
+		}
+
+		printk(KERN_ERR "no space left, need %llu, %llu delalloc bytes"
+		       ", %llu bytes_used, %llu bytes_reserved, "
+		       "%llu bytes_pinned, %llu bytes_readonly, %llu may use"
+		       "%llu total\n", bytes, data_sinfo->bytes_delalloc,
+		       data_sinfo->bytes_used, data_sinfo->bytes_reserved,
+		       data_sinfo->bytes_pinned, data_sinfo->bytes_readonly,
+		       data_sinfo->bytes_may_use, data_sinfo->total_bytes);
+		return -ENOSPC;
+	}
+	data_sinfo->bytes_may_use += bytes;
+	BTRFS_I(inode)->reserved_bytes += bytes;
+	spin_unlock(&data_sinfo->lock);
+
+	return btrfs_check_metadata_free_space(root);
+}
+
+/*
+ * if there was an error for whatever reason after calling
+ * btrfs_check_data_free_space, call this so we can cleanup the counters.
+ */
+void btrfs_free_reserved_data_space(struct btrfs_root *root,
+				    struct inode *inode, u64 bytes)
+{
+	struct btrfs_space_info *data_sinfo;
+
+	/* make sure bytes are sectorsize aligned */
+	bytes = (bytes + root->sectorsize - 1) & ~((u64)root->sectorsize - 1);
+
+	data_sinfo = BTRFS_I(inode)->space_info;
+	spin_lock(&data_sinfo->lock);
+	data_sinfo->bytes_may_use -= bytes;
+	BTRFS_I(inode)->reserved_bytes -= bytes;
+	spin_unlock(&data_sinfo->lock);
+}
+
+/* called when we are adding a delalloc extent to the inode's io_tree */
+void btrfs_delalloc_reserve_space(struct btrfs_root *root, struct inode *inode,
+				  u64 bytes)
+{
+	struct btrfs_space_info *data_sinfo;
+
+	/* get the space info for where this inode will be storing its data */
+	data_sinfo = BTRFS_I(inode)->space_info;
+
+	/* make sure we have enough space to handle the data first */
+	spin_lock(&data_sinfo->lock);
+	data_sinfo->bytes_delalloc += bytes;
+
+	/*
+	 * we are adding a delalloc extent without calling
+	 * btrfs_check_data_free_space first.  This happens on a weird
+	 * writepage condition, but shouldn't hurt our accounting
+	 */
+	if (unlikely(bytes > BTRFS_I(inode)->reserved_bytes)) {
+		data_sinfo->bytes_may_use -= BTRFS_I(inode)->reserved_bytes;
+		BTRFS_I(inode)->reserved_bytes = 0;
+	} else {
+		data_sinfo->bytes_may_use -= bytes;
+		BTRFS_I(inode)->reserved_bytes -= bytes;
+	}
+
+	spin_unlock(&data_sinfo->lock);
+}
+
+/* called when we are clearing an delalloc extent from the inode's io_tree */
+void btrfs_delalloc_free_space(struct btrfs_root *root, struct inode *inode,
+			      u64 bytes)
+{
+	struct btrfs_space_info *info;
+
+	info = BTRFS_I(inode)->space_info;
+
+	spin_lock(&info->lock);
+	info->bytes_delalloc -= bytes;
+	spin_unlock(&info->lock);
+}
+
 static int do_chunk_alloc(struct btrfs_trans_handle *trans,
 			  struct btrfs_root *extent_root, u64 alloc_bytes,
 			  u64 flags, int force)
@@ -3105,6 +3337,10 @@ static void dump_space_info(struct btrfs_space_info *info, u64 bytes)
 	       (unsigned long long)(info->total_bytes - info->bytes_used -
 				    info->bytes_pinned - info->bytes_reserved),
 	       (info->full) ? "" : "not ");
+	printk(KERN_INFO "space_info total=%llu, pinned=%llu, delalloc=%llu,"
+	       " may_use=%llu, used=%llu\n", info->total_bytes,
+	       info->bytes_pinned, info->bytes_delalloc, info->bytes_may_use,
+	       info->bytes_used);
 
 	down_read(&info->groups_sem);
 	list_for_each_entry(cache, &info->block_groups, list) {
@@ -3131,24 +3367,10 @@ static int __btrfs_reserve_extent(struct btrfs_trans_handle *trans,
 {
 	int ret;
 	u64 search_start = 0;
-	u64 alloc_profile;
 	struct btrfs_fs_info *info = root->fs_info;
 
-	if (data) {
-		alloc_profile = info->avail_data_alloc_bits &
-			info->data_alloc_profile;
-		data = BTRFS_BLOCK_GROUP_DATA | alloc_profile;
-	} else if (root == root->fs_info->chunk_root) {
-		alloc_profile = info->avail_system_alloc_bits &
-			info->system_alloc_profile;
-		data = BTRFS_BLOCK_GROUP_SYSTEM | alloc_profile;
-	} else {
-		alloc_profile = info->avail_metadata_alloc_bits &
-			info->metadata_alloc_profile;
-		data = BTRFS_BLOCK_GROUP_METADATA | alloc_profile;
-	}
+	data = btrfs_get_alloc_profile(root, data);
 again:
-	data = btrfs_reduce_alloc_profile(root, data);
 	/*
 	 * the only place that sets empty_size is btrfs_realloc_node, which
 	 * is not called recursively on allocations
diff --git a/fs/btrfs/file.c b/fs/btrfs/file.c
index 872f104576e5..dc78954861b3 100644
--- a/fs/btrfs/file.c
+++ b/fs/btrfs/file.c
@@ -1091,19 +1091,24 @@ static ssize_t btrfs_file_write(struct file *file, const char __user *buf,
 		WARN_ON(num_pages > nrptrs);
 		memset(pages, 0, sizeof(struct page *) * nrptrs);
 
-		ret = btrfs_check_free_space(root, write_bytes, 0);
+		ret = btrfs_check_data_free_space(root, inode, write_bytes);
 		if (ret)
 			goto out;
 
 		ret = prepare_pages(root, file, pages, num_pages,
 				    pos, first_index, last_index,
 				    write_bytes);
-		if (ret)
+		if (ret) {
+			btrfs_free_reserved_data_space(root, inode,
+						       write_bytes);
 			goto out;
+		}
 
 		ret = btrfs_copy_from_user(pos, num_pages,
 					   write_bytes, pages, buf);
 		if (ret) {
+			btrfs_free_reserved_data_space(root, inode,
+						       write_bytes);
 			btrfs_drop_pages(pages, num_pages);
 			goto out;
 		}
@@ -1111,8 +1116,11 @@ static ssize_t btrfs_file_write(struct file *file, const char __user *buf,
 		ret = dirty_and_release_pages(NULL, root, file, pages,
 					      num_pages, pos, write_bytes);
 		btrfs_drop_pages(pages, num_pages);
-		if (ret)
+		if (ret) {
+			btrfs_free_reserved_data_space(root, inode,
+						       write_bytes);
 			goto out;
+		}
 
 		if (will_write) {
 			btrfs_fdatawrite_range(inode->i_mapping, pos,
@@ -1136,6 +1144,8 @@ static ssize_t btrfs_file_write(struct file *file, const char __user *buf,
 	}
 out:
 	mutex_unlock(&inode->i_mutex);
+	if (ret)
+		err = ret;
 
 out_nolock:
 	kfree(pages);
diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c
index 3cee77ae03c8..7d4f948bc22a 100644
--- a/fs/btrfs/inode.c
+++ b/fs/btrfs/inode.c
@@ -102,34 +102,6 @@ static int btrfs_init_inode_security(struct inode *inode,  struct inode *dir)
 }
 
 /*
- * a very lame attempt at stopping writes when the FS is 85% full.  There
- * are countless ways this is incorrect, but it is better than nothing.
- */
-int btrfs_check_free_space(struct btrfs_root *root, u64 num_required,
-			   int for_del)
-{
-	u64 total;
-	u64 used;
-	u64 thresh;
-	int ret = 0;
-
-	spin_lock(&root->fs_info->delalloc_lock);
-	total = btrfs_super_total_bytes(&root->fs_info->super_copy);
-	used = btrfs_super_bytes_used(&root->fs_info->super_copy);
-	if (for_del)
-		thresh = total * 90;
-	else
-		thresh = total * 85;
-
-	do_div(thresh, 100);
-
-	if (used + root->fs_info->delalloc_bytes + num_required > thresh)
-		ret = -ENOSPC;
-	spin_unlock(&root->fs_info->delalloc_lock);
-	return ret;
-}
-
-/*
  * this does all the hard work for inserting an inline extent into
  * the btree.  The caller should have done a btrfs_drop_extents so that
  * no overlapping inline items exist in the btree
@@ -1190,6 +1162,7 @@ static int btrfs_set_bit_hook(struct inode *inode, u64 start, u64 end,
 	 */
 	if (!(old & EXTENT_DELALLOC) && (bits & EXTENT_DELALLOC)) {
 		struct btrfs_root *root = BTRFS_I(inode)->root;
+		btrfs_delalloc_reserve_space(root, inode, end - start + 1);
 		spin_lock(&root->fs_info->delalloc_lock);
 		BTRFS_I(inode)->delalloc_bytes += end - start + 1;
 		root->fs_info->delalloc_bytes += end - start + 1;
@@ -1223,9 +1196,12 @@ static int btrfs_clear_bit_hook(struct inode *inode, u64 start, u64 end,
 			       (unsigned long long)end - start + 1,
 			       (unsigned long long)
 			       root->fs_info->delalloc_bytes);
+			btrfs_delalloc_free_space(root, inode, (u64)-1);
 			root->fs_info->delalloc_bytes = 0;
 			BTRFS_I(inode)->delalloc_bytes = 0;
 		} else {
+			btrfs_delalloc_free_space(root, inode,
+						  end - start + 1);
 			root->fs_info->delalloc_bytes -= end - start + 1;
 			BTRFS_I(inode)->delalloc_bytes -= end - start + 1;
 		}
@@ -2245,10 +2221,6 @@ static int btrfs_unlink(struct inode *dir, struct dentry *dentry)
 
 	root = BTRFS_I(dir)->root;
 
-	ret = btrfs_check_free_space(root, 1, 1);
-	if (ret)
-		goto fail;
-
 	trans = btrfs_start_transaction(root, 1);
 
 	btrfs_set_trans_block_group(trans, dir);
@@ -2261,7 +2233,6 @@ static int btrfs_unlink(struct inode *dir, struct dentry *dentry)
 	nr = trans->blocks_used;
 
 	btrfs_end_transaction_throttle(trans, root);
-fail:
 	btrfs_btree_balance_dirty(root, nr);
 	return ret;
 }
@@ -2284,10 +2255,6 @@ static int btrfs_rmdir(struct inode *dir, struct dentry *dentry)
 		return -ENOTEMPTY;
 	}
 
-	ret = btrfs_check_free_space(root, 1, 1);
-	if (ret)
-		goto fail;
-
 	trans = btrfs_start_transaction(root, 1);
 	btrfs_set_trans_block_group(trans, dir);
 
@@ -2304,7 +2271,6 @@ static int btrfs_rmdir(struct inode *dir, struct dentry *dentry)
 fail_trans:
 	nr = trans->blocks_used;
 	ret = btrfs_end_transaction_throttle(trans, root);
-fail:
 	btrfs_btree_balance_dirty(root, nr);
 
 	if (ret && !err)
@@ -2818,7 +2784,7 @@ int btrfs_cont_expand(struct inode *inode, loff_t size)
 	if (size <= hole_start)
 		return 0;
 
-	err = btrfs_check_free_space(root, 1, 0);
+	err = btrfs_check_metadata_free_space(root);
 	if (err)
 		return err;
 
@@ -3014,6 +2980,7 @@ static noinline void init_btrfs_i(struct inode *inode)
 	bi->last_trans = 0;
 	bi->logged_trans = 0;
 	bi->delalloc_bytes = 0;
+	bi->reserved_bytes = 0;
 	bi->disk_i_size = 0;
 	bi->flags = 0;
 	bi->index_cnt = (u64)-1;
@@ -3035,6 +3002,7 @@ static int btrfs_init_locked_inode(struct inode *inode, void *p)
 	inode->i_ino = args->ino;
 	init_btrfs_i(inode);
 	BTRFS_I(inode)->root = args->root;
+	btrfs_set_inode_space_info(args->root, inode);
 	return 0;
 }
 
@@ -3455,6 +3423,7 @@ static struct inode *btrfs_new_inode(struct btrfs_trans_handle *trans,
 	BTRFS_I(inode)->index_cnt = 2;
 	BTRFS_I(inode)->root = root;
 	BTRFS_I(inode)->generation = trans->transid;
+	btrfs_set_inode_space_info(root, inode);
 
 	if (mode & S_IFDIR)
 		owner = 0;
@@ -3602,7 +3571,7 @@ static int btrfs_mknod(struct inode *dir, struct dentry *dentry,
 	if (!new_valid_dev(rdev))
 		return -EINVAL;
 
-	err = btrfs_check_free_space(root, 1, 0);
+	err = btrfs_check_metadata_free_space(root);
 	if (err)
 		goto fail;
 
@@ -3665,7 +3634,7 @@ static int btrfs_create(struct inode *dir, struct dentry *dentry,
 	u64 objectid;
 	u64 index = 0;
 
-	err = btrfs_check_free_space(root, 1, 0);
+	err = btrfs_check_metadata_free_space(root);
 	if (err)
 		goto fail;
 	trans = btrfs_start_transaction(root, 1);
@@ -3733,7 +3702,7 @@ static int btrfs_link(struct dentry *old_dentry, struct inode *dir,
 		return -ENOENT;
 
 	btrfs_inc_nlink(inode);
-	err = btrfs_check_free_space(root, 1, 0);
+	err = btrfs_check_metadata_free_space(root);
 	if (err)
 		goto fail;
 	err = btrfs_set_inode_index(dir, &index);
@@ -3779,7 +3748,7 @@ static int btrfs_mkdir(struct inode *dir, struct dentry *dentry, int mode)
 	u64 index = 0;
 	unsigned long nr = 1;
 
-	err = btrfs_check_free_space(root, 1, 0);
+	err = btrfs_check_metadata_free_space(root);
 	if (err)
 		goto out_unlock;
 
@@ -4336,7 +4305,7 @@ int btrfs_page_mkwrite(struct vm_area_struct *vma, struct page *page)
 	u64 page_start;
 	u64 page_end;
 
-	ret = btrfs_check_free_space(root, PAGE_CACHE_SIZE, 0);
+	ret = btrfs_check_data_free_space(root, inode, PAGE_CACHE_SIZE);
 	if (ret)
 		goto out;
 
@@ -4349,6 +4318,7 @@ again:
 
 	if ((page->mapping != inode->i_mapping) ||
 	    (page_start >= size)) {
+		btrfs_free_reserved_data_space(root, inode, PAGE_CACHE_SIZE);
 		/* page got truncated out from underneath us */
 		goto out_unlock;
 	}
@@ -4631,7 +4601,7 @@ static int btrfs_rename(struct inode *old_dir, struct dentry *old_dentry,
 	if (old_inode->i_ino == BTRFS_FIRST_FREE_OBJECTID)
 		return -EXDEV;
 
-	ret = btrfs_check_free_space(root, 1, 0);
+	ret = btrfs_check_metadata_free_space(root);
 	if (ret)
 		goto out_unlock;
 
@@ -4749,7 +4719,7 @@ static int btrfs_symlink(struct inode *dir, struct dentry *dentry,
 	if (name_len > BTRFS_MAX_INLINE_DATA_SIZE(root))
 		return -ENAMETOOLONG;
 
-	err = btrfs_check_free_space(root, 1, 0);
+	err = btrfs_check_metadata_free_space(root);
 	if (err)
 		goto out_fail;
 
diff --git a/fs/btrfs/ioctl.c b/fs/btrfs/ioctl.c
index 988fdc8b49eb..bca729fc80c8 100644
--- a/fs/btrfs/ioctl.c
+++ b/fs/btrfs/ioctl.c
@@ -70,7 +70,7 @@ static noinline int create_subvol(struct btrfs_root *root,
 	u64 index = 0;
 	unsigned long nr = 1;
 
-	ret = btrfs_check_free_space(root, 1, 0);
+	ret = btrfs_check_metadata_free_space(root);
 	if (ret)
 		goto fail_commit;
 
@@ -203,7 +203,7 @@ static int create_snapshot(struct btrfs_root *root, struct dentry *dentry,
 	if (!root->ref_cows)
 		return -EINVAL;
 
-	ret = btrfs_check_free_space(root, 1, 0);
+	ret = btrfs_check_metadata_free_space(root);
 	if (ret)
 		goto fail_unlock;
 
@@ -374,7 +374,7 @@ static int btrfs_defrag_file(struct file *file)
 	unsigned long i;
 	int ret;
 
-	ret = btrfs_check_free_space(root, inode->i_size, 0);
+	ret = btrfs_check_data_free_space(root, inode, inode->i_size);
 	if (ret)
 		return -ENOSPC;
 
diff --git a/fs/compat_ioctl.c b/fs/compat_ioctl.c
index 39bd4d38e889..45e59d3c7f1f 100644
--- a/fs/compat_ioctl.c
+++ b/fs/compat_ioctl.c
@@ -1913,6 +1913,9 @@ COMPATIBLE_IOCTL(FIONREAD)  /* This is also TIOCINQ */
 /* 0x00 */
 COMPATIBLE_IOCTL(FIBMAP)
 COMPATIBLE_IOCTL(FIGETBSZ)
+/* 'X' - originally XFS but some now in the VFS */
+COMPATIBLE_IOCTL(FIFREEZE)
+COMPATIBLE_IOCTL(FITHAW)
 /* RAID */
 COMPATIBLE_IOCTL(RAID_VERSION)
 COMPATIBLE_IOCTL(GET_ARRAY_INFO)
diff --git a/fs/dcache.c b/fs/dcache.c
index 937df0fb0da5..07e2d4a44bda 100644
--- a/fs/dcache.c
+++ b/fs/dcache.c
@@ -1180,7 +1180,7 @@ struct dentry *d_obtain_alias(struct inode *inode)
 	iput(inode);
 	return res;
 }
-EXPORT_SYMBOL_GPL(d_obtain_alias);
+EXPORT_SYMBOL(d_obtain_alias);
 
 /**
  * d_splice_alias - splice a disconnected dentry into the tree if one exists
diff --git a/fs/ext4/balloc.c b/fs/ext4/balloc.c
index 9a50b8052dcf..de9459b4cb94 100644
--- a/fs/ext4/balloc.c
+++ b/fs/ext4/balloc.c
@@ -609,7 +609,9 @@ int ext4_claim_free_blocks(struct ext4_sb_info *sbi,
  */
 int ext4_should_retry_alloc(struct super_block *sb, int *retries)
 {
-	if (!ext4_has_free_blocks(EXT4_SB(sb), 1) || (*retries)++ > 3)
+	if (!ext4_has_free_blocks(EXT4_SB(sb), 1) ||
+	    (*retries)++ > 3 ||
+	    !EXT4_SB(sb)->s_journal)
 		return 0;
 
 	jbd_debug(1, "%s: retrying operation after ENOSPC\n", sb->s_id);
diff --git a/fs/ext4/inode.c b/fs/ext4/inode.c
index 51cdd13e1c31..c7fed5b18745 100644
--- a/fs/ext4/inode.c
+++ b/fs/ext4/inode.c
@@ -2544,7 +2544,7 @@ retry:
 
 		ext4_journal_stop(handle);
 
-		if (mpd.retval == -ENOSPC) {
+		if ((mpd.retval == -ENOSPC) && sbi->s_journal) {
 			/* commit the transaction which would
 			 * free blocks released in the transaction
 			 * and try again
diff --git a/fs/ext4/super.c b/fs/ext4/super.c
index a5732c58f676..39d1993cfa13 100644
--- a/fs/ext4/super.c
+++ b/fs/ext4/super.c
@@ -3091,7 +3091,6 @@ static int ext4_freeze(struct super_block *sb)
 
 		/* Journal blocked and flushed, clear needs_recovery flag. */
 		EXT4_CLEAR_INCOMPAT_FEATURE(sb, EXT4_FEATURE_INCOMPAT_RECOVER);
-		ext4_commit_super(sb, EXT4_SB(sb)->s_es, 1);
 		error = ext4_commit_super(sb, EXT4_SB(sb)->s_es, 1);
 		if (error)
 			goto out;
diff --git a/fs/jffs2/background.c b/fs/jffs2/background.c
index 3cceef4ad2b7..e9580104b6ba 100644
--- a/fs/jffs2/background.c
+++ b/fs/jffs2/background.c
@@ -95,13 +95,17 @@ static int jffs2_garbage_collect_thread(void *_c)
 			spin_unlock(&c->erase_completion_lock);
 			
 
-		/* This thread is purely an optimisation. But if it runs when
-		   other things could be running, it actually makes things a
-		   lot worse. Use yield() and put it at the back of the runqueue
-		   every time. Especially during boot, pulling an inode in
-		   with read_inode() is much preferable to having the GC thread
-		   get there first. */
-		yield();
+		/* Problem - immediately after bootup, the GCD spends a lot
+		 * of time in places like jffs2_kill_fragtree(); so much so
+		 * that userspace processes (like gdm and X) are starved
+		 * despite plenty of cond_resched()s and renicing.  Yield()
+		 * doesn't help, either (presumably because userspace and GCD
+		 * are generally competing for a higher latency resource -
+		 * disk).
+		 * This forces the GCD to slow the hell down.   Pulling an
+		 * inode in with read_inode() is much preferable to having
+		 * the GC thread get there first. */
+		schedule_timeout_interruptible(msecs_to_jiffies(50));
 
 		/* Put_super will send a SIGKILL and then wait on the sem.
 		 */
diff --git a/fs/jffs2/readinode.c b/fs/jffs2/readinode.c
index 6ca08ad887c0..1fc1e92356ee 100644
--- a/fs/jffs2/readinode.c
+++ b/fs/jffs2/readinode.c
@@ -220,7 +220,7 @@ static int jffs2_add_tn_to_tree(struct jffs2_sb_info *c,
 				struct jffs2_tmp_dnode_info *tn)
 {
 	uint32_t fn_end = tn->fn->ofs + tn->fn->size;
-	struct jffs2_tmp_dnode_info *this;
+	struct jffs2_tmp_dnode_info *this, *ptn;
 
 	dbg_readinode("insert fragment %#04x-%#04x, ver %u at %08x\n", tn->fn->ofs, fn_end, tn->version, ref_offset(tn->fn->raw));
 
@@ -251,11 +251,18 @@ static int jffs2_add_tn_to_tree(struct jffs2_sb_info *c,
 	if (this) {
 		/* If the node is coincident with another at a lower address,
 		   back up until the other node is found. It may be relevant */
-		while (this->overlapped)
-			this = tn_prev(this);
-
-		/* First node should never be marked overlapped */
-		BUG_ON(!this);
+		while (this->overlapped) {
+			ptn = tn_prev(this);
+			if (!ptn) {
+				/*
+				 * We killed a node which set the overlapped
+				 * flags during the scan. Fix it up.
+				 */
+				this->overlapped = 0;
+				break;
+			}
+			this = ptn;
+		}
 		dbg_readinode("'this' found %#04x-%#04x (%s)\n", this->fn->ofs, this->fn->ofs + this->fn->size, this->fn ? "data" : "hole");
 	}
 
@@ -360,7 +367,17 @@ static int jffs2_add_tn_to_tree(struct jffs2_sb_info *c,
 			}
 			if (!this->overlapped)
 				break;
-			this = tn_prev(this);
+
+			ptn = tn_prev(this);
+			if (!ptn) {
+				/*
+				 * We killed a node which set the overlapped
+				 * flags during the scan. Fix it up.
+				 */
+				this->overlapped = 0;
+				break;
+			}
+			this = ptn;
 		}
 	}
 
@@ -456,8 +473,15 @@ static int jffs2_build_inode_fragtree(struct jffs2_sb_info *c,
 		eat_last(&rii->tn_root, &last->rb);
 		ver_insert(&ver_root, last);
 
-		if (unlikely(last->overlapped))
-			continue;
+		if (unlikely(last->overlapped)) {
+			if (pen)
+				continue;
+			/*
+			 * We killed a node which set the overlapped
+			 * flags during the scan. Fix it up.
+			 */
+			last->overlapped = 0;
+		}
 
 		/* Now we have a bunch of nodes in reverse version
 		   order, in the tree at ver_root. Most of the time,
diff --git a/fs/ocfs2/alloc.c b/fs/ocfs2/alloc.c
index 60fe74035db5..3a9e5deed74d 100644
--- a/fs/ocfs2/alloc.c
+++ b/fs/ocfs2/alloc.c
@@ -4796,6 +4796,29 @@ out:
 	return ret;
 }
 
+static int ocfs2_replace_extent_rec(struct inode *inode,
+				    handle_t *handle,
+				    struct ocfs2_path *path,
+				    struct ocfs2_extent_list *el,
+				    int split_index,
+				    struct ocfs2_extent_rec *split_rec)
+{
+	int ret;
+
+	ret = ocfs2_path_bh_journal_access(handle, inode, path,
+					   path_num_items(path) - 1);
+	if (ret) {
+		mlog_errno(ret);
+		goto out;
+	}
+
+	el->l_recs[split_index] = *split_rec;
+
+	ocfs2_journal_dirty(handle, path_leaf_bh(path));
+out:
+	return ret;
+}
+
 /*
  * Mark part or all of the extent record at split_index in the leaf
  * pointed to by path as written. This removes the unwritten
@@ -4885,7 +4908,9 @@ static int __ocfs2_mark_extent_written(struct inode *inode,
 
 	if (ctxt.c_contig_type == CONTIG_NONE) {
 		if (ctxt.c_split_covers_rec)
-			el->l_recs[split_index] = *split_rec;
+			ret = ocfs2_replace_extent_rec(inode, handle,
+						       path, el,
+						       split_index, split_rec);
 		else
 			ret = ocfs2_split_and_insert(inode, handle, path, et,
 						     &last_eb_bh, split_index,
diff --git a/fs/ocfs2/dlm/dlmmaster.c b/fs/ocfs2/dlm/dlmmaster.c
index 54e182a27caf..0a2813947853 100644
--- a/fs/ocfs2/dlm/dlmmaster.c
+++ b/fs/ocfs2/dlm/dlmmaster.c
@@ -1849,12 +1849,12 @@ int dlm_assert_master_handler(struct o2net_msg *msg, u32 len, void *data,
 		if (!mle) {
 			if (res->owner != DLM_LOCK_RES_OWNER_UNKNOWN &&
 			    res->owner != assert->node_idx) {
-				mlog(ML_ERROR, "assert_master from "
-					  "%u, but current owner is "
-					  "%u! (%.*s)\n",
-				       assert->node_idx, res->owner,
-				       namelen, name);
-				goto kill;
+				mlog(ML_ERROR, "DIE! Mastery assert from %u, "
+				     "but current owner is %u! (%.*s)\n",
+				     assert->node_idx, res->owner, namelen,
+				     name);
+				__dlm_print_one_lock_resource(res);
+				BUG();
 			}
 		} else if (mle->type != DLM_MLE_MIGRATION) {
 			if (res->owner != DLM_LOCK_RES_OWNER_UNKNOWN) {
diff --git a/fs/ocfs2/dlm/dlmthread.c b/fs/ocfs2/dlm/dlmthread.c
index d1295203029f..4060bb328bc8 100644
--- a/fs/ocfs2/dlm/dlmthread.c
+++ b/fs/ocfs2/dlm/dlmthread.c
@@ -181,8 +181,7 @@ static int dlm_purge_lockres(struct dlm_ctxt *dlm,
 
 		spin_lock(&res->spinlock);
 		/* This ensures that clear refmap is sent after the set */
-		__dlm_wait_on_lockres_flags(res, (DLM_LOCK_RES_SETREF_INPROG |
-						  DLM_LOCK_RES_MIGRATING));
+		__dlm_wait_on_lockres_flags(res, DLM_LOCK_RES_SETREF_INPROG);
 		spin_unlock(&res->spinlock);
 
 		/* clear our bit from the master's refmap, ignore errors */
diff --git a/fs/ocfs2/dlm/dlmunlock.c b/fs/ocfs2/dlm/dlmunlock.c
index 86ca085ef324..fcf879ed6930 100644
--- a/fs/ocfs2/dlm/dlmunlock.c
+++ b/fs/ocfs2/dlm/dlmunlock.c
@@ -117,11 +117,11 @@ static enum dlm_status dlmunlock_common(struct dlm_ctxt *dlm,
 	else
 		BUG_ON(res->owner == dlm->node_num);
 
-	spin_lock(&dlm->spinlock);
+	spin_lock(&dlm->ast_lock);
 	/* We want to be sure that we're not freeing a lock
 	 * that still has AST's pending... */
 	in_use = !list_empty(&lock->ast_list);
-	spin_unlock(&dlm->spinlock);
+	spin_unlock(&dlm->ast_lock);
 	if (in_use) {
 	       mlog(ML_ERROR, "lockres %.*s: Someone is calling dlmunlock "
 		    "while waiting for an ast!", res->lockname.len,
diff --git a/fs/ocfs2/dlmglue.c b/fs/ocfs2/dlmglue.c
index 206a2370876a..7219a86d34cc 100644
--- a/fs/ocfs2/dlmglue.c
+++ b/fs/ocfs2/dlmglue.c
@@ -320,9 +320,14 @@ static void ocfs2_schedule_blocked_lock(struct ocfs2_super *osb,
 					struct ocfs2_lock_res *lockres);
 static inline void ocfs2_recover_from_dlm_error(struct ocfs2_lock_res *lockres,
 						int convert);
-#define ocfs2_log_dlm_error(_func, _err, _lockres) do {			\
-	mlog(ML_ERROR, "DLM error %d while calling %s on resource %s\n", \
-	     _err, _func, _lockres->l_name);				\
+#define ocfs2_log_dlm_error(_func, _err, _lockres) do {					\
+	if ((_lockres)->l_type != OCFS2_LOCK_TYPE_DENTRY)				\
+		mlog(ML_ERROR, "DLM error %d while calling %s on resource %s\n",	\
+		     _err, _func, _lockres->l_name);					\
+	else										\
+		mlog(ML_ERROR, "DLM error %d while calling %s on resource %.*s%08x\n",	\
+		     _err, _func, OCFS2_DENTRY_LOCK_INO_START - 1, (_lockres)->l_name,	\
+		     (unsigned int)ocfs2_get_dentry_lock_ino(_lockres));		\
 } while (0)
 static int ocfs2_downconvert_thread(void *arg);
 static void ocfs2_downconvert_on_unlock(struct ocfs2_super *osb,
diff --git a/fs/ocfs2/ocfs2.h b/fs/ocfs2/ocfs2.h
index 077384135f4e..946d3c34b90b 100644
--- a/fs/ocfs2/ocfs2.h
+++ b/fs/ocfs2/ocfs2.h
@@ -341,6 +341,9 @@ struct ocfs2_super
 	struct ocfs2_node_map		osb_recovering_orphan_dirs;
 	unsigned int			*osb_orphan_wipes;
 	wait_queue_head_t		osb_wipe_event;
+
+	/* used to protect metaecc calculation check of xattr. */
+	spinlock_t osb_xattr_lock;
 };
 
 #define OCFS2_SB(sb)	    ((struct ocfs2_super *)(sb)->s_fs_info)
diff --git a/fs/ocfs2/super.c b/fs/ocfs2/super.c
index b1cb38fbe807..7ac83a81ee55 100644
--- a/fs/ocfs2/super.c
+++ b/fs/ocfs2/super.c
@@ -1537,6 +1537,13 @@ static int ocfs2_get_sector(struct super_block *sb,
 	unlock_buffer(*bh);
 	ll_rw_block(READ, 1, bh);
 	wait_on_buffer(*bh);
+	if (!buffer_uptodate(*bh)) {
+		mlog_errno(-EIO);
+		brelse(*bh);
+		*bh = NULL;
+		return -EIO;
+	}
+
 	return 0;
 }
 
@@ -1747,6 +1754,7 @@ static int ocfs2_initialize_super(struct super_block *sb,
 	INIT_LIST_HEAD(&osb->blocked_lock_list);
 	osb->blocked_lock_count = 0;
 	spin_lock_init(&osb->osb_lock);
+	spin_lock_init(&osb->osb_xattr_lock);
 	ocfs2_init_inode_steal_slot(osb);
 
 	atomic_set(&osb->alloc_stats.moves, 0);
diff --git a/fs/ocfs2/xattr.c b/fs/ocfs2/xattr.c
index 915039fffe6e..4ddd788add67 100644
--- a/fs/ocfs2/xattr.c
+++ b/fs/ocfs2/xattr.c
@@ -82,13 +82,14 @@ struct ocfs2_xattr_set_ctxt {
 
 #define OCFS2_XATTR_ROOT_SIZE	(sizeof(struct ocfs2_xattr_def_value_root))
 #define OCFS2_XATTR_INLINE_SIZE	80
+#define OCFS2_XATTR_HEADER_GAP	4
 #define OCFS2_XATTR_FREE_IN_IBODY	(OCFS2_MIN_XATTR_INLINE_SIZE \
 					 - sizeof(struct ocfs2_xattr_header) \
-					 - sizeof(__u32))
+					 - OCFS2_XATTR_HEADER_GAP)
 #define OCFS2_XATTR_FREE_IN_BLOCK(ptr)	((ptr)->i_sb->s_blocksize \
 					 - sizeof(struct ocfs2_xattr_block) \
 					 - sizeof(struct ocfs2_xattr_header) \
-					 - sizeof(__u32))
+					 - OCFS2_XATTR_HEADER_GAP)
 
 static struct ocfs2_xattr_def_value_root def_xv = {
 	.xv.xr_list.l_count = cpu_to_le16(1),
@@ -274,10 +275,12 @@ static int ocfs2_read_xattr_bucket(struct ocfs2_xattr_bucket *bucket,
 			       bucket->bu_blocks, bucket->bu_bhs, 0,
 			       NULL);
 	if (!rc) {
+		spin_lock(&OCFS2_SB(bucket->bu_inode->i_sb)->osb_xattr_lock);
 		rc = ocfs2_validate_meta_ecc_bhs(bucket->bu_inode->i_sb,
 						 bucket->bu_bhs,
 						 bucket->bu_blocks,
 						 &bucket_xh(bucket)->xh_check);
+		spin_unlock(&OCFS2_SB(bucket->bu_inode->i_sb)->osb_xattr_lock);
 		if (rc)
 			mlog_errno(rc);
 	}
@@ -310,9 +313,11 @@ static void ocfs2_xattr_bucket_journal_dirty(handle_t *handle,
 {
 	int i;
 
+	spin_lock(&OCFS2_SB(bucket->bu_inode->i_sb)->osb_xattr_lock);
 	ocfs2_compute_meta_ecc_bhs(bucket->bu_inode->i_sb,
 				   bucket->bu_bhs, bucket->bu_blocks,
 				   &bucket_xh(bucket)->xh_check);
+	spin_unlock(&OCFS2_SB(bucket->bu_inode->i_sb)->osb_xattr_lock);
 
 	for (i = 0; i < bucket->bu_blocks; i++)
 		ocfs2_journal_dirty(handle, bucket->bu_bhs[i]);
@@ -1507,7 +1512,7 @@ static int ocfs2_xattr_set_entry(struct inode *inode,
 		last += 1;
 	}
 
-	free = min_offs - ((void *)last - xs->base) - sizeof(__u32);
+	free = min_offs - ((void *)last - xs->base) - OCFS2_XATTR_HEADER_GAP;
 	if (free < 0)
 		return -EIO;
 
@@ -2190,7 +2195,7 @@ static int ocfs2_xattr_can_be_in_inode(struct inode *inode,
 		last += 1;
 	}
 
-	free = min_offs - ((void *)last - xs->base) - sizeof(__u32);
+	free = min_offs - ((void *)last - xs->base) - OCFS2_XATTR_HEADER_GAP;
 	if (free < 0)
 		return 0;
 
@@ -2592,8 +2597,9 @@ static int __ocfs2_xattr_set_handle(struct inode *inode,
 
 	if (!ret) {
 		/* Update inode ctime. */
-		ret = ocfs2_journal_access(ctxt->handle, inode, xis->inode_bh,
-					   OCFS2_JOURNAL_ACCESS_WRITE);
+		ret = ocfs2_journal_access_di(ctxt->handle, inode,
+					      xis->inode_bh,
+					      OCFS2_JOURNAL_ACCESS_WRITE);
 		if (ret) {
 			mlog_errno(ret);
 			goto out;
@@ -5060,8 +5066,8 @@ try_again:
 	xh_free_start = le16_to_cpu(xh->xh_free_start);
 	header_size = sizeof(struct ocfs2_xattr_header) +
 			count * sizeof(struct ocfs2_xattr_entry);
-	max_free = OCFS2_XATTR_BUCKET_SIZE -
-		le16_to_cpu(xh->xh_name_value_len) - header_size;
+	max_free = OCFS2_XATTR_BUCKET_SIZE - header_size -
+		le16_to_cpu(xh->xh_name_value_len) - OCFS2_XATTR_HEADER_GAP;
 
 	mlog_bug_on_msg(header_size > blocksize, "bucket %llu has header size "
 			"of %u which exceed block size\n",
@@ -5094,7 +5100,7 @@ try_again:
 			need = 0;
 	}
 
-	free = xh_free_start - header_size;
+	free = xh_free_start - header_size - OCFS2_XATTR_HEADER_GAP;
 	/*
 	 * We need to make sure the new name/value pair
 	 * can exist in the same block.
@@ -5127,7 +5133,8 @@ try_again:
 			}
 
 			xh_free_start = le16_to_cpu(xh->xh_free_start);
-			free = xh_free_start - header_size;
+			free = xh_free_start - header_size
+				- OCFS2_XATTR_HEADER_GAP;
 			if (xh_free_start % blocksize < need)
 				free -= xh_free_start % blocksize;