From: Jeff Layton Subject: [PATCH 2/2] vfs: force reval on dentry of bind mounted files for LOOKUP_OPEN Date: Tue, 10 Nov 2009 11:27:36 -0500 Message-ID: <1257870456-31188-3-git-send-email-jlayton@redhat.com> References: <1257870456-31188-1-git-send-email-jlayton@redhat.com> Cc: ebiederm@xmission.com, adobriyan@gmail.com, viro@ZenIV.linux.org.uk, jamie@shareable.org To: linux-kernel@vger.kernel.org, linux-fsdevel@vger.kernel.org, linux-nfs@vger.kernel.org Return-path: In-Reply-To: <1257870456-31188-1-git-send-email-jlayton@redhat.com> Sender: linux-fsdevel-owner@vger.kernel.org List-ID: In the case of a bind mounted file, the path walking code will assume that a cached dentry is valid and doesn't revalidate it. This is a problem for NFSv4 in a way that's similar to LAST_BIND symlinks. Fix this by revalidating the dentry if LOOKUP_OPEN is set and __follow_mount returns true. Signed-off-by: Jeff Layton --- fs/namei.c | 9 +++++++-- 1 files changed, 7 insertions(+), 2 deletions(-) diff --git a/fs/namei.c b/fs/namei.c index 5c8ef80..b7c9747 100644 --- a/fs/namei.c +++ b/fs/namei.c @@ -839,6 +839,7 @@ static int do_lookup(struct nameidata *nd, struct qstr *name, { struct vfsmount *mnt = nd->path.mnt; struct dentry *dentry = __d_lookup(nd->path.dentry, name); + int error = 0; if (!dentry) goto need_lookup; @@ -847,8 +848,9 @@ static int do_lookup(struct nameidata *nd, struct qstr *name, done: path->mnt = mnt; path->dentry = dentry; - __follow_mount(path); - return 0; + if (__follow_mount(path)) + error = open_reval_path(path, nd); + return error; need_lookup: dentry = real_lookup(nd->path.dentry, name, nd); @@ -1836,6 +1838,9 @@ do_last: error = -ELOOP; if (flag & O_NOFOLLOW) goto exit_dput; + error = open_reval_path(&path, &nd); + if (error) + goto exit; } error = -ENOENT; -- 1.5.5.6