From mboxrd@z Thu Jan 1 00:00:00 1970 From: Waiman Long Subject: [PATCH 4/4] dcache: don't need to take d_lock in prepend_path() Date: Tue, 19 Feb 2013 13:50:59 -0500 Message-ID: <1361299859-27056-5-git-send-email-Waiman.Long@hp.com> References: <1361299859-27056-1-git-send-email-Waiman.Long@hp.com> Cc: Waiman Long , linux-kernel@vger.kernel.org To: linux-fsdevel@vger.kernel.org, Alexander Viro Return-path: In-Reply-To: <1361299859-27056-1-git-send-email-Waiman.Long@hp.com> Sender: linux-kernel-owner@vger.kernel.org List-Id: linux-fsdevel.vger.kernel.org The d_lock was used in prepend_path() to protect dentry->d_name from being changed under the hood. As the caller of prepend_path() has to take the rename_lock before calling into it, there is no chance that d_name will be changed. The d_lock lock is only needed when the rename_lock is not taken. Signed-off-by: Waiman Long --- fs/dcache.c | 3 +-- 1 files changed, 1 insertions(+), 2 deletions(-) diff --git a/fs/dcache.c b/fs/dcache.c index b1487e2..0e911fc 100644 --- a/fs/dcache.c +++ b/fs/dcache.c @@ -2547,6 +2547,7 @@ static int prepend_name(char **buffer, int *buflen, struct qstr *name) * @buflen: pointer to buffer length * * Caller holds the rename_lock. + * There is no need to lock the dentry as its name cannot be changed. */ static int prepend_path(const struct path *path, const struct path *root, @@ -2573,9 +2574,7 @@ static int prepend_path(const struct path *path, } parent = dentry->d_parent; prefetch(parent); - spin_lock(&dentry->d_lock); error = prepend_name(buffer, buflen, &dentry->d_name); - spin_unlock(&dentry->d_lock); if (!error) error = prepend(buffer, buflen, "/", 1); if (error) -- 1.7.1