From mboxrd@z Thu Jan 1 00:00:00 1970 From: Al Viro Subject: Re: [RFC][PATCH] debugfs: Fix corrupted loop in debugfs_remove_recursive Date: Thu, 12 Jun 2014 11:43:14 +0100 Message-ID: <20140612104314.GB18016@ZenIV.linux.org.uk> References: <20140609140607.3e3aae39@gandalf.local.home> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Cc: Greg Kroah-Hartman , LKML , linux-fsdevel@vger.kernel.org, Andrew Morton To: Steven Rostedt Return-path: Received: from zeniv.linux.org.uk ([195.92.253.2]:57092 "EHLO ZenIV.linux.org.uk" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S932916AbaFLKnR (ORCPT ); Thu, 12 Jun 2014 06:43:17 -0400 Content-Disposition: inline In-Reply-To: <20140609140607.3e3aae39@gandalf.local.home> Sender: linux-fsdevel-owner@vger.kernel.org List-ID: On Mon, Jun 09, 2014 at 02:06:07PM -0400, Steven Rostedt wrote: > When these are called, the d_entry and inode locks need to be released > to call the instance creation and deletion code. That code has its own > accounting and locking to serialize everything to prevent multiple > users from causing harm. As the parent "instance" directory can not > be modified this simplifies things. Yecchhh... Looking at debugfs: static inline int debugfs_positive(struct dentry *dentry) { return dentry->d_inode && !d_unhashed(dentry); } ... if (debugfs_positive(dentry)) { if (dentry->d_inode) { What the hell? parent = dentry->d_parent; if (!parent || !parent->d_inode) return; Huh? First of all, ->d_parent is *never* NULL. Moreover, it can't be a negative dentry. What's more, if debugfs_rename() is ever used for cross-directory renames, this tree-walker is buggered - it'll happily walk up "back" into a directory it has never visited...