[PATCH] mnt_expire is protected by namespace_sem, no need for vfsmount_lock
Al Viro [Sat, 22 Mar 2008 20:14:30 +0000 (16:14 -0400)]
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>

fs/namespace.c

index 7bd74b2..94f026e 100644 (file)
@@ -262,10 +262,8 @@ static struct vfsmount *clone_mnt(struct vfsmount *old, struct dentry *root,
                /* stick the duplicate mount on the same expiry list
                 * as the original if that was on one */
                if (flag & CL_EXPIRE) {
-                       spin_lock(&vfsmount_lock);
                        if (!list_empty(&old->mnt_expire))
                                list_add(&mnt->mnt_expire, &old->mnt_expire);
-                       spin_unlock(&vfsmount_lock);
                }
        }
        return mnt;
@@ -1127,11 +1125,9 @@ static noinline int do_move_mount(struct nameidata *nd, char *old_name)
        if (err)
                goto out1;
 
-       spin_lock(&vfsmount_lock);
        /* if the mount is moved, it should no longer be expire
         * automatically */
        list_del_init(&old_nd.path.mnt->mnt_expire);
-       spin_unlock(&vfsmount_lock);
 out1:
        mutex_unlock(&nd->path.dentry->d_inode->i_mutex);
 out:
@@ -1198,12 +1194,9 @@ int do_add_mount(struct vfsmount *newmnt, struct nameidata *nd,
        if ((err = graft_tree(newmnt, nd)))
                goto unlock;
 
-       if (fslist) {
-               /* add to the specified expiration list */
-               spin_lock(&vfsmount_lock);
+       if (fslist) /* add to the specified expiration list */
                list_add_tail(&newmnt->mnt_expire, fslist);
-               spin_unlock(&vfsmount_lock);
-       }
+
        up_write(&namespace_sem);
        return 0;