From mboxrd@z Thu Jan 1 00:00:00 1970 From: Alan Cox Subject: Re: [PATCH v2 1/1] gpio: gpio-wcove: Fix GPIO control register offset calculation Date: Thu, 29 Jun 2017 20:28:57 +0100 Message-ID: <20170629202857.359bec9d@alans-desktop> References: <4b57167184fa9aa4d1c183d4c8df6e6a600a7dbd.1498087915.git.sathyanarayanan.kuppuswamy@linux.intel.com> Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Return-path: Received: from www.llwyncelyn.cymru ([82.70.14.225]:60706 "EHLO fuzix.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751988AbdF2T3P (ORCPT ); Thu, 29 Jun 2017 15:29:15 -0400 In-Reply-To: Sender: linux-gpio-owner@vger.kernel.org List-Id: linux-gpio@vger.kernel.org To: Linus Walleij Cc: Kuppuswamy Sathyanarayanan , Mika Westerberg , Andy Shevchenko , Hans de Goede , "Rafael J. Wysocki" , "linux-gpio@vger.kernel.org" , "linux-kernel@vger.kernel.org" , Sathyanarayanan Kuppuswamy Natarajan , Graeme Gregory > So where can I get a handle on the people inside Intel who are obviously > using ACPI GPIO class for shoehorning what we in the linux kernel call > syscon or register bit misc access into the GPIO ACPI container just > because they feel it is convenient? It's a Windowsism and since Windows is the primary OS shipped the vendors of client platforms do what is needed to make Windows work nicely. > They need to invent a NEW ACPI four-character thing and call that > "misc register bit" (_MRB?) or whatever and have it bind to syscon. > This is not working. Short of Microsoft adopting such a standard I don't think it would make any difference (beyond making life worse because you'd have a new _MRB that wasn't used by Windows so nobody ever tested). > It feels like I am starting to maintain Intel's swiss army knife for misc > register manipulation, and that should not be done by "virtual GPIO" > because just look at it: It's an ACPIism not an Intelism. I expect it's there on other vendors devices and the same things will pop up as Windows/ARM platforms with ACPI appear. > General-purpose input/output - yeah that sounds like something > going in/out of the system right? It's become an ACPI interface for controlling all sorts of system state in a way that works nicely in Windows. Rightly or wrongly that's the situation and we are still the tail not the dog. Alan