From mboxrd@z Thu Jan 1 00:00:00 1970 From: stigge@antcom.de (Roland Stigge) Date: Mon, 17 Dec 2012 14:51:58 +0100 Subject: [PATCH RESEND 0/6 v10] gpio: Add block GPIO In-Reply-To: <50CF1EF1.2070601@antcom.de> References: <1355495185-24220-1-git-send-email-stigge@antcom.de> <50CB68AB.5070806@grandegger.com> <50CBBB25.20002@antcom.de> <50CF03FB.2030100@grandegger.com> <50CF0744.7040404@grandegger.com> <50CF1EF1.2070601@antcom.de> Message-ID: <50CF237E.5020409@antcom.de> To: linux-arm-kernel@lists.infradead.org List-Id: linux-arm-kernel.lists.infradead.org Hi Wolfgang, On 12/17/2012 02:32 PM, Roland Stigge wrote: > And I guess Russell is right: If possible, we should write outputs > simultaneously via ODSR (plus OWER/OWDR/OWSR) instead of separate set/clear. > > I wonder if we need to save/restore the state of OWSR at every write > operation or if we need/can cache it. Assuming that block GPIO are the > only code in the kernel that manipulates ODSR. Can you please test the following: +static void at91_gpiolib_set_block(struct gpio_chip *chip, unsigned long mask, unsigned long val) +{ + struct at91_gpio_chip *at91_gpio = to_at91_gpio_chip(chip); + void __iomem *pio = at91_gpio->regbase; + + __raw_writel(~mask, pio + PIO_OWDR); + __raw_writel(mask, pio + PIO_OWER); + __raw_writel(val, pio + PIO_ODSR); +} Would caching OWSR be a significant speedup here? Thanks in advance, Roland From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1753109Ab2LQNwD (ORCPT ); Mon, 17 Dec 2012 08:52:03 -0500 Received: from antcom.de ([188.40.178.216]:47018 "EHLO chuck.antcom.de" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753078Ab2LQNwB (ORCPT ); Mon, 17 Dec 2012 08:52:01 -0500 Message-ID: <50CF237E.5020409@antcom.de> Date: Mon, 17 Dec 2012 14:51:58 +0100 From: Roland Stigge Organization: ANTCOM IT Research & Development User-Agent: Mozilla/5.0 (X11; Linux i686; rv:10.0.11) Gecko/20121123 Icedove/10.0.11 MIME-Version: 1.0 To: Wolfgang Grandegger CC: rmallon@gmail.com, gregkh@linuxfoundation.org, linus.walleij@linaro.org, broonie@opensource.wolfsonmicro.com, linux-kernel@vger.kernel.org, w.sang@pengutronix.de, grant.likely@secretlab.ca, daniel-gl@gmx.net, sr@denx.de, plagnioj@jcrosoft.com, linux-arm-kernel@lists.infradead.org, highguy@gmail.com Subject: Re: [PATCH RESEND 0/6 v10] gpio: Add block GPIO References: <1355495185-24220-1-git-send-email-stigge@antcom.de> <50CB68AB.5070806@grandegger.com> <50CBBB25.20002@antcom.de> <50CF03FB.2030100@grandegger.com> <50CF0744.7040404@grandegger.com> <50CF1EF1.2070601@antcom.de> In-Reply-To: <50CF1EF1.2070601@antcom.de> X-Enigmail-Version: 1.4 OpenPGP: url=subkeys.pgp.net Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 7bit Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Hi Wolfgang, On 12/17/2012 02:32 PM, Roland Stigge wrote: > And I guess Russell is right: If possible, we should write outputs > simultaneously via ODSR (plus OWER/OWDR/OWSR) instead of separate set/clear. > > I wonder if we need to save/restore the state of OWSR at every write > operation or if we need/can cache it. Assuming that block GPIO are the > only code in the kernel that manipulates ODSR. Can you please test the following: +static void at91_gpiolib_set_block(struct gpio_chip *chip, unsigned long mask, unsigned long val) +{ + struct at91_gpio_chip *at91_gpio = to_at91_gpio_chip(chip); + void __iomem *pio = at91_gpio->regbase; + + __raw_writel(~mask, pio + PIO_OWDR); + __raw_writel(mask, pio + PIO_OWER); + __raw_writel(val, pio + PIO_ODSR); +} Would caching OWSR be a significant speedup here? Thanks in advance, Roland