Merge branch 'upstream-linus' of git://git.kernel.org/pub/scm/linux/k…

…ernel/git/mfasheh/ocfs2

* 'upstream-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mfasheh/ocfs2: (28 commits)
  ocfs2: Teach ocfs2_drop_lock() to use ->set_lvb() callback
  ocfs2: Remove ->unblock lockres operation
  ocfs2: move downconvert worker to lockres ops
  ocfs2: Remove unused dlmglue functions
  ocfs2: Have the metadata lock use generic dlmglue functions
  ocfs2: Add ->set_lvb callback in dlmglue
  ocfs2: Add ->check_downconvert callback in dlmglue
  ocfs2: Check for refreshing locks in generic unblock function
  ocfs2: don't unconditionally pass LVB flags
  ocfs2: combine inode and generic blocking AST functions
  ocfs2: Add ->get_osb() dlmglue locking operation
  ocfs2: remove ->unlock_ast() callback from ocfs2_lock_res_ops
  ocfs2: combine inode and generic AST functions
  ocfs2: Clean up lock resource refresh flags
  ocfs2: Remove i_generation from inode lock names
  ocfs2: Encode i_generation in the meta data lvb
  ocfs2: Free up some space in the lvb
  ocfs2: Remove special casing for inode creation in ocfs2_dentry_attach_lock()
  ocfs2: manually d_move() during ocfs2_rename()
  [PATCH] Allow file systems to manually d_move() inside of ->rename()
  ...

fs/namei.c

@@ -2370,7 +2370,8 @@ static int vfs_rename_dir(struct inode *old_dir, struct dentry *old_dentry,
 		dput(new_dentry);
 	}
 	if (!error)
-		d_move(old_dentry,new_dentry);
+		if (!(old_dir->i_sb->s_type->fs_flags & FS_RENAME_DOES_D_MOVE))
+			d_move(old_dentry,new_dentry);
 	return error;
 }
 
@@ -2393,8 +2394,7 @@ static int vfs_rename_other(struct inode *old_dir, struct dentry *old_dentry,
 	else
 		error = old_dir->i_op->rename(old_dir, old_dentry, new_dir, new_dentry);
 	if (!error) {
-		/* The following d_move() should become unconditional */
-		if (!(old_dir->i_sb->s_type->fs_flags & FS_ODD_RENAME))
+		if (!(old_dir->i_sb->s_type->fs_flags & FS_RENAME_DOES_D_MOVE))
 			d_move(old_dentry, new_dentry);
 	}
 	if (target)

fs/nfs/dir.c

@@ -1669,8 +1669,7 @@ static int nfs_rename(struct inode *old_dir, struct dentry *old_dentry,
 	if (rehash)
 		d_rehash(rehash);
 	if (!error) {
-		if (!S_ISDIR(old_inode->i_mode))
-			d_move(old_dentry, new_dentry);
+		d_move(old_dentry, new_dentry);
 		nfs_renew_times(new_dentry);
 		nfs_set_verifier(new_dentry, nfs_save_change_attribute(new_dir));
 	}

fs/nfs/super.c

@@ -71,15 +71,15 @@ static struct file_system_type nfs_fs_type = {
 	.name		= "nfs",
 	.get_sb		= nfs_get_sb,
 	.kill_sb	= nfs_kill_super,
-	.fs_flags	= FS_ODD_RENAME|FS_REVAL_DOT|FS_BINARY_MOUNTDATA,
+	.fs_flags	= FS_RENAME_DOES_D_MOVE|FS_REVAL_DOT|FS_BINARY_MOUNTDATA,
 };
 
 struct file_system_type nfs_xdev_fs_type = {
 	.owner		= THIS_MODULE,
 	.name		= "nfs",
 	.get_sb		= nfs_xdev_get_sb,
 	.kill_sb	= nfs_kill_super,
-	.fs_flags	= FS_ODD_RENAME|FS_REVAL_DOT|FS_BINARY_MOUNTDATA,
+	.fs_flags	= FS_RENAME_DOES_D_MOVE|FS_REVAL_DOT|FS_BINARY_MOUNTDATA,
 };
 
 static struct super_operations nfs_sops = {
@@ -107,23 +107,23 @@ static struct file_system_type nfs4_fs_type = {
 	.name		= "nfs4",
 	.get_sb		= nfs4_get_sb,
 	.kill_sb	= nfs4_kill_super,
-	.fs_flags	= FS_ODD_RENAME|FS_REVAL_DOT|FS_BINARY_MOUNTDATA,
+	.fs_flags	= FS_RENAME_DOES_D_MOVE|FS_REVAL_DOT|FS_BINARY_MOUNTDATA,
 };
 
 struct file_system_type nfs4_xdev_fs_type = {
 	.owner		= THIS_MODULE,
 	.name		= "nfs4",
 	.get_sb		= nfs4_xdev_get_sb,
 	.kill_sb	= nfs4_kill_super,
-	.fs_flags	= FS_ODD_RENAME|FS_REVAL_DOT|FS_BINARY_MOUNTDATA,
+	.fs_flags	= FS_RENAME_DOES_D_MOVE|FS_REVAL_DOT|FS_BINARY_MOUNTDATA,
 };
 
 struct file_system_type nfs4_referral_fs_type = {
 	.owner		= THIS_MODULE,
 	.name		= "nfs4",
 	.get_sb		= nfs4_referral_get_sb,
 	.kill_sb	= nfs4_kill_super,
-	.fs_flags	= FS_ODD_RENAME|FS_REVAL_DOT|FS_BINARY_MOUNTDATA,
+	.fs_flags	= FS_RENAME_DOES_D_MOVE|FS_REVAL_DOT|FS_BINARY_MOUNTDATA,
 };
 
 static struct super_operations nfs4_sops = {

fs/ocfs2/cluster/tcp_internal.h

@@ -44,11 +44,17 @@
  * locking semantics of the file system using the protocol.  It should 
  * be somewhere else, I'm sure, but right now it isn't.
  *
+ * New in version 4:
+ * 	- Remove i_generation from lock names for better stat performance.
+ *
+ * New in version 3:
+ * 	- Replace dentry votes with a cluster lock
+ *
  * New in version 2:
  * 	- full 64 bit i_size in the metadata lock lvbs
  * 	- introduction of "rw" lock and pushing meta/data locking down
  */
-#define O2NET_PROTOCOL_VERSION 2ULL
+#define O2NET_PROTOCOL_VERSION 4ULL
 struct o2net_handshake {
 	__be64	protocol_version;
 	__be64	connector_id;