Merge tag 'fs.fixes.v6.0-rc3' of git://git.kernel.org/pub/scm/linux/kernel/git/vfs/idmapping

Pull file_remove_privs() fix from Christian Brauner:
 "As part of Stefan's and Jens' work to add async buffered write
  support to xfs we refactored file_remove_privs() and added
  __file_remove_privs() to avoid calling __remove_privs() when
  IOCB_NOWAIT is passed.

  While debugging a recent performance regression report I found that
  during review we missed that commit faf99b563558 ("fs: add
  __remove_file_privs() with flags parameter") accidently changed
  behavior when dentry_needs_remove_privs() returns zero.

  Before the commit it would still call inode_has_no_xattr() setting
  the S_NOSEC bit and thereby avoiding even calling into
  dentry_needs_remove_privs() the next time this function is called.
  After that commit inode_has_no_xattr() would only be called if
  __remove_privs() had to be called.

  Restore the old behavior. This is likely the cause of the performance
  regression"

* tag 'fs.fixes.v6.0-rc3' of git://git.kernel.org/pub/scm/linux/kernel/git/vfs/idmapping:
  fs: __file_remove_privs(): restore call to inode_has_no_xattr()

1 files changed, 8 insertions, 6 deletions

diff --git a/fs/inode.c b/fs/inode.c
index 6462276dfdf0..ba1de23c13c1 100644
--- a/fs/inode.c
+++ b/fs/inode.c
@@ -2018,23 +2018,25 @@ static int __file_remove_privs(struct file *file, unsigned int flags)
 {
 	struct dentry *dentry = file_dentry(file);
 	struct inode *inode = file_inode(file);
-	int error;
+	int error = 0;
 	int kill;
 
 	if (IS_NOSEC(inode) || !S_ISREG(inode->i_mode))
 		return 0;
 
 	kill = dentry_needs_remove_privs(dentry);
-	if (kill <= 0)
+	if (kill < 0)
 		return kill;
 
-	if (flags & IOCB_NOWAIT)
-		return -EAGAIN;
+	if (kill) {
+		if (flags & IOCB_NOWAIT)
+			return -EAGAIN;
+
+		error = __remove_privs(file_mnt_user_ns(file), dentry, kill);
+	}
 
-	error = __remove_privs(file_mnt_user_ns(file), dentry, kill);
 	if (!error)
 		inode_has_no_xattr(inode);
-
 	return error;
 }