From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S964892AbWGPE5i (ORCPT ); Sun, 16 Jul 2006 00:57:38 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S964893AbWGPE5i (ORCPT ); Sun, 16 Jul 2006 00:57:38 -0400 Received: from mail3.sea5.speakeasy.net ([69.17.117.5]:21642 "EHLO mail3.sea5.speakeasy.net") by vger.kernel.org with ESMTP id S964892AbWGPE5i (ORCPT ); Sun, 16 Jul 2006 00:57:38 -0400 Date: Sun, 16 Jul 2006 00:57:34 -0400 (EDT) From: James Morris X-X-Sender: jmorris@d.namei To: Catherine Zhang cc: linux-kernel@vger.kernel.org, michal.k.k.piotrowski@gmail.com, catalin.marinas@gmail.com, czhang.us@gmail.com, Stephen Smalley Subject: Re: kernel memory leak fix for af_unix datagram getpeersec In-Reply-To: <20060716045731.GA303@jiayuguan.watson.ibm.com> Message-ID: References: <20060716045731.GA303@jiayuguan.watson.ibm.com> MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org On Sun, 16 Jul 2006, Catherine Zhang wrote: > > Hi, Catalin and Michal, > > Enclosed please find the patch against 2.6.18-rc1 that fixed the kernel > memory leak problem. The core kernel code doesn't know anything about secdata, as it is up to the LSM to determine what it is (in the case of SELinux, a kmalloc'd buffer). So, you need a cleanup hook which allows the LSM to free the secdata if required. - James -- James Morris