aufs: possible bugfix, uncached acl

When a branch filesystem doesn't cache ACL, aufs should not cache
either.  Until now aufs has never met such fs, but theoretically it
could happen.  Actually, in linux-v5.1-rc1, NFSv3 changed its behaviour
by the commit
	ded52fbe7020 2019-02-20 nfs: fix xfstest generic/099 failed on nfsv3
The commit ded52fbe7020 doesn't "forget" the previous acl though.
Doesn't it mean that the obsoleted acl is kept until NFS's attribute
cache is expired?  I don't know.  I've asked it on LKML, but got no
answer.
Signed-off-by: J. R. Okajima <hooanon05g@gmail.com>
(cherry picked from commit e448daa00228186b869356fdef8d98d9f95caf53)

fs/aufs/posix_acl.c

@@ -37,7 +37,9 @@ struct posix_acl *aufs_get_acl(struct inode *inode, int type)
 
 	/* always topmost only */
 	acl = get_acl(h_inode, type);
-	if (!IS_ERR_OR_NULL(acl))
+	if (IS_ERR(acl))
+		forget_cached_acl(inode, type);
+	else
 		set_cached_acl(inode, type, acl);
 
 out:
@@ -78,12 +80,12 @@ int aufs_set_acl(struct inode *inode, struct posix_acl *acl, int type)
 	}
 
 	ssz = au_sxattr(dentry, inode, &arg);
+	/* forget even it if succeeds since the branch might set differently */
+	forget_cached_acl(inode, type);
 	dput(dentry);
 	err = ssz;
-	if (ssz >= 0) {
+	if (ssz >= 0)
 		err = 0;
-		set_cached_acl(inode, type, acl);
-	}
 
 out:
 	return err;