From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S264331AbTDPMMZ (for ); Wed, 16 Apr 2003 08:12:25 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S264333AbTDPMMZ (for ); Wed, 16 Apr 2003 08:12:25 -0400 Received: from pc2-cwma1-4-cust86.swan.cable.ntl.com ([213.105.254.86]:58818 "EHLO lxorguk.ukuu.org.uk") by vger.kernel.org with ESMTP id S264331AbTDPMMY (for ); Wed, 16 Apr 2003 08:12:24 -0400 Subject: Re: [RFC] Port 0x80 writes considered harmful From: Alan Cox To: Kasper Dupont Cc: Linux Kernel Mailing List In-Reply-To: <3E9CF81F.F7CCCADF@daimi.au.dk> References: <3E9CF81F.F7CCCADF@daimi.au.dk> Content-Type: text/plain Content-Transfer-Encoding: 7bit Organization: Message-Id: <1050492360.28591.36.camel@dhcp22.swansea.linux.org.uk> Mime-Version: 1.0 X-Mailer: Ximian Evolution 1.2.2 (1.2.2-5) Date: 16 Apr 2003 12:26:01 +0100 Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org On Mer, 2003-04-16 at 07:28, Kasper Dupont wrote: > Then I inserted port 0x80 writes in the loop, and at that > point there would be parallel port output during the entire > execution of the loop. > > What would be the best solution to my problem? If the parallel port is picking up 0x80 writes then either its at a very weird and very stupid address or its got some kind of decode problem.