[PATCH 4/4] vfs: Call d_dname from dentry_path

[ebiederm-aS9lmoZGLiVWk0Htik3J/w@public.gmane.org (Eric W. Biederman)]

If the root dentry implements d_dname instead of ignoring it in
__dentry_path call d_dname and return the result.

The motivating example are files under /proc/<pid>/ns/ are bind
mounted into other locations.  This change allow the name of the
dentry that is bind mounted to show up in /proc/<pid>/mountinfo.
However in general this is correct behavior any time dentry_path is
called on a dentry that implements d_dname.

Signed-off-by: "Eric W. Biederman" <ebiederm-aS9lmoZGLiVWk0Htik3J/w@public.gmane.org>
---
 fs/dcache.c |    2 ++
 1 files changed, 2 insertions(+), 0 deletions(-)

diff --git a/fs/dcache.c b/fs/dcache.c
index c5c7847ff84b..a7a925fb3ce7 100644
--- a/fs/dcache.c
+++ b/fs/dcache.c
@@ -3162,6 +3162,8 @@ restart:
 	done_seqretry(&rename_lock, seq);
 	if (error)
 		goto Elong;
+	if (dentry->d_op && dentry->d_op->d_dname)
+		retval = dentry->d_op->d_dname(dentry, buf, len);
 	return retval;
 Elong:
 	return ERR_PTR(-ENAMETOOLONG);
-- 
1.7.5.4