From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1752485AbXCLVOb (ORCPT ); Mon, 12 Mar 2007 17:14:31 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1752547AbXCLVOb (ORCPT ); Mon, 12 Mar 2007 17:14:31 -0400 Received: from mga02.intel.com ([134.134.136.20]:24971 "EHLO mga02.intel.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752526AbXCLVO3 (ORCPT ); Mon, 12 Mar 2007 17:14:29 -0400 X-ExtLoop1: 1 X-IronPort-AV: i="4.14,275,1170662400"; d="scan'208"; a="208317924:sNHT18793278" Date: Mon, 12 Mar 2007 14:11:26 -0700 From: Kristen Carlson Accardi To: Alan Cox Cc: Jeff Garzik , len.brown@intel.com, Linux Kernel Mailing List , IDE/ATA development list , Andrew Morton , Linus Torvalds Subject: Re: libata-acpi: allow _GTF on SATA, but disable on PATA for now Message-Id: <20070312141126.651cf967.kristen.c.accardi@intel.com> In-Reply-To: <20070312220015.094b7759@lxorguk.ukuu.org.uk> References: <200703100559.l2A5x6kP017578@hera.kernel.org> <45F296B9.7000005@garzik.org> <20070310164059.4ba68bcc@lxorguk.ukuu.org.uk> <20070312095528.3dd7a693.kristen.c.accardi@intel.com> <20070312220015.094b7759@lxorguk.ukuu.org.uk> X-Mailer: Sylpheed version 2.2.10 (GTK+ 2.10.8; i386-redhat-linux-gnu) Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org On Mon, 12 Mar 2007 22:00:15 +0000 Alan Cox wrote: > > Meanwhile, to solve the problem for these people, and without introducing > > a lot of new code at this stage in the release, perhaps you can suggest > > a more accurate way to detect PATA vs. SATA so we can fix the bug? > > I think the problem starts with the question. It supposes the BIOS knows > the answer to the question in the first place. It also supposes we > haven't reconfigured the port in some cases I believe. > > My guess is the problem needs approaching from the other end. Ask the > question "What does the ACPI BIOS expect it to be ?". If I follow the > spec correctly we can answer that question by asking what methods (_SDD, > _GTF etc) it provides. > > I don't think anything else is reliable. > > Alan > Ah - ok, so you are saying we could query the device to see if it supports STM/GTM and then if we were to use the workaround I suggested for this bug we could just refuse to execute GTF on devices which support STM (since they may require execution of STM first). If you guys decide you want the workaround rather than just either disable acpi or remove the acpi patch entirely, let me know and I'd be glad to rework it, otherwise I'm assuming you have decided to just disable acpi all together.