From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1751597AbWHAMgm (ORCPT ); Tue, 1 Aug 2006 08:36:42 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1751590AbWHAMgl (ORCPT ); Tue, 1 Aug 2006 08:36:41 -0400 Received: from mail.aknet.ru ([82.179.72.26]:17415 "EHLO mail.aknet.ru") by vger.kernel.org with ESMTP id S1751538AbWHAMgl (ORCPT ); Tue, 1 Aug 2006 08:36:41 -0400 Message-ID: <44CF4B84.6010409@aknet.ru> Date: Tue, 01 Aug 2006 16:39:32 +0400 From: Stas Sergeev User-Agent: Thunderbird 1.5.0.4 (X11/20060614) MIME-Version: 1.0 To: Chuck Ebbert <76306.1226@compuserve.com> Cc: "akpm@osdl.org" , zach@vmware.com, rohitseth@google.com, JBeulich@novell.com, ak@muc.de, linux-kernel Subject: Re: + espfix-code-cleanup.patch added to -mm tree References: <200607312226_MC3-1-C6A9-1757@compuserve.com> In-Reply-To: <200607312226_MC3-1-C6A9-1757@compuserve.com> Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org Hi. Chuck Ebbert wrote: > we are on a ring0 32-bit stack that's not zero-based. If an exception > occurs in that state, UNWIND_ESPFIX_STACK restores the proper kernel > SS and ESP but on return from the exception nothing restores the espfix > stack. I guess this isn't a problem now because exceptions in kernel > mode are fatal but a kernel debugger might have problems here? Perhaps you are right, but... unless there is some quick way to mark that part of code "undebuggable", I'll better leave that for the debugger maintainers to think about.