From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1752072AbXBWAQN (ORCPT ); Thu, 22 Feb 2007 19:16:13 -0500 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1752070AbXBWAQN (ORCPT ); Thu, 22 Feb 2007 19:16:13 -0500 Received: from shawidc-mo1.cg.shawcable.net ([24.71.223.10]:53619 "EHLO pd3mo2so.prod.shaw.ca" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752072AbXBWAQL (ORCPT ); Thu, 22 Feb 2007 19:16:11 -0500 Date: Thu, 22 Feb 2007 18:15:03 -0600 From: Robert Hancock Subject: Re: [PATCH] ACPI driver support for pata In-reply-to: <20070222182939.56d43a68@lxorguk.ukuu.org.uk> To: Alan Cc: Jeff Garzik , akpm@osdl.org, linux-kernel@vger.kernel.org Message-id: <45DE3207.6020707@shaw.ca> MIME-version: 1.0 Content-type: text/plain; charset=ISO-8859-1; format=flowed Content-transfer-encoding: 7bit References: <45DD2E12.6000705@shaw.ca> <45DD3F20.7050907@pobox.com> <20070222142439.50cff3e9@lxorguk.ukuu.org.uk> <45DDCEC4.7050208@garzik.org> <20070222182939.56d43a68@lxorguk.ukuu.org.uk> User-Agent: Thunderbird 1.5.0.9 (Windows/20061207) Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org Alan wrote: > On Thu, 22 Feb 2007 12:11:32 -0500 > Jeff Garzik wrote: > >> Alan wrote: >>> ACPI is the only way to do cable handling on the Nvidia PATA chipset. The >> You failed to quote the salient part of the message. Disliking a >> separate pata_acpi driver in no way invalidates your statement (quoted >> above). > > If you drive a device by the ACPI interface you don't get to fiddle with > it directly or you end up in a murky world of undefined and ungood > behaviour. Testing and vendor information both say pata_acpi is the right > way to drive Nvidia PATA ports. Couldn't be do this generically inside libata core somehow, i.e. try to use ACPI to set the proper mode and fall back to the driver-specific mode setting code if that didn't work? I think if we could do that it would solve a number of problems (i.e. we could prevent it from doing this on SATA controllers which appear to be IDE based on the PCI ID, like the NVIDIA SATA controllers, since the _GTM and _STM methods seem to have undefined behavior on SATA). This would also eliminate the need for mkinitrd, people, etc. to know that they're supposed to be loading this other pata_acpi driver instead, since any PATA driver could take advantage of this feature. -- Robert Hancock Saskatoon, SK, Canada To email, remove "nospam" from hancockr@nospamshaw.ca Home Page: http://www.roberthancock.com/