From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1752585AbbGNGOG (ORCPT ); Tue, 14 Jul 2015 02:14:06 -0400 Received: from metis.ext.pengutronix.de ([92.198.50.35]:38887 "EHLO metis.ext.pengutronix.de" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751858AbbGNGOF (ORCPT ); Tue, 14 Jul 2015 02:14:05 -0400 Date: Tue, 14 Jul 2015 08:13:59 +0200 From: Sascha Hauer To: Ludovic Desroches Cc: linux-arm-kernel@lists.infradead.org, linux-gpio@vger.kernel.org, devicetree@vger.kernel.org, linux-kernel@vger.kernel.org, swarren@wwwdotorg.org, linus.walleij@linaro.org, nicolas.ferre@atmel.com Subject: Re: [RESEND PATCH 2/2] pinctrl: introduce complex pin description Message-ID: <20150714061359.GC5161@pengutronix.de> References: <1433948699-19800-1-git-send-email-ludovic.desroches@atmel.com> <1433948699-19800-3-git-send-email-ludovic.desroches@atmel.com> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <1433948699-19800-3-git-send-email-ludovic.desroches@atmel.com> X-Sent-From: Pengutronix Hildesheim X-URL: http://www.pengutronix.de/ X-IRC: #ptxdist @freenode X-Accept-Language: de,en X-Accept-Content-Type: text/plain X-Uptime: 07:59:35 up 19 days, 21 min, 54 users, load average: 0.13, 0.13, 0.14 User-Agent: Mutt/1.5.21 (2010-09-15) X-SA-Exim-Connect-IP: 2001:67c:670:100:1d::c0 X-SA-Exim-Mail-From: sha@pengutronix.de X-SA-Exim-Scanned: No (on metis.ext.pengutronix.de); SAEximRunCond expanded to false X-PTX-Original-Recipient: linux-kernel@vger.kernel.org Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On Wed, Jun 10, 2015 at 05:04:57PM +0200, Ludovic Desroches wrote: > Using a string to describe a pin in the device tree can be not enough. > Some controllers may need extra information to fully describe a pin. It > concerns mainly controllers which have a per pin muxing approach which > don't fit well the notions of groups and functions. > Instead of using a pin name, a 32 bit value is used. The 16 least > significant bits are used for the pin number. Other 16 bits can be used to > store extra parameters. In the Mediatek driver we use 'pinmux' as name for the property containing the combined pin number / mux value defines. 'pinmux' better describes what it is... > + > + if (pctldesc->complex_pin_desc) > + ret = of_property_count_u32_elems(np, "pins"); > + else > + ret = of_property_count_strings(np, "pins"); ... and has the advantage that you don't have to pass in a complex_pin_desc variable from the driver as the different property name inherently carries this information. 'pins' can then stay a property containing only strings. > if (ret < 0) { > ret = of_property_count_strings(np, "groups"); > if (ret < 0) > @@ -297,11 +305,12 @@ int pinconf_generic_dt_subnode_to_map(struct pinctrl_dev *pctldev, > if (type == PIN_MAP_TYPE_INVALID) > type = PIN_MAP_TYPE_CONFIGS_GROUP; > subnode_target_type = "groups"; > + pins_prop = false; > } else { > if (type == PIN_MAP_TYPE_INVALID) > type = PIN_MAP_TYPE_CONFIGS_PIN; > } > - strings_count = ret; > + items_count = ret; > > ret = of_property_read_string(np, "function", &function); > if (ret < 0) { > @@ -326,17 +335,31 @@ int pinconf_generic_dt_subnode_to_map(struct pinctrl_dev *pctldev, > if (num_configs) > reserve++; > > - reserve *= strings_count; > + reserve *= items_count; > > ret = pinctrl_utils_reserve_map(pctldev, map, reserved_maps, > num_maps, reserve); > if (ret < 0) > goto exit; > > - of_property_for_each_string(np, subnode_target_type, prop, group) { > + items_name = kmalloc_array(items_count, sizeof(char *), GFP_KERNEL); > + if (!items_name) > + goto exit; > + if (pctldesc->complex_pin_desc && pins_prop) { > + of_property_for_each_u32(np, subnode_target_type, prop, cur, val) { > + pin_id = val & PINCTRL_PIN_MASK; > + items_name[i++] = pctldesc->pins[pin_id].name; > + } I don't like that even though pins have numbers here they are converted to strings which the driver later has to search in a list just to convert it back into the number. This is quite inefficient. I guess this could be optimized later, but it would be nice to have the pin number directly in the driver. Sascha -- Pengutronix e.K. | | Industrial Linux Solutions | http://www.pengutronix.de/ | Peiner Str. 6-8, 31137 Hildesheim, Germany | Phone: +49-5121-206917-0 | Amtsgericht Hildesheim, HRA 2686 | Fax: +49-5121-206917-5555 | From mboxrd@z Thu Jan 1 00:00:00 1970 From: s.hauer@pengutronix.de (Sascha Hauer) Date: Tue, 14 Jul 2015 08:13:59 +0200 Subject: [RESEND PATCH 2/2] pinctrl: introduce complex pin description In-Reply-To: <1433948699-19800-3-git-send-email-ludovic.desroches@atmel.com> References: <1433948699-19800-1-git-send-email-ludovic.desroches@atmel.com> <1433948699-19800-3-git-send-email-ludovic.desroches@atmel.com> Message-ID: <20150714061359.GC5161@pengutronix.de> To: linux-arm-kernel@lists.infradead.org List-Id: linux-arm-kernel.lists.infradead.org On Wed, Jun 10, 2015 at 05:04:57PM +0200, Ludovic Desroches wrote: > Using a string to describe a pin in the device tree can be not enough. > Some controllers may need extra information to fully describe a pin. It > concerns mainly controllers which have a per pin muxing approach which > don't fit well the notions of groups and functions. > Instead of using a pin name, a 32 bit value is used. The 16 least > significant bits are used for the pin number. Other 16 bits can be used to > store extra parameters. In the Mediatek driver we use 'pinmux' as name for the property containing the combined pin number / mux value defines. 'pinmux' better describes what it is... > + > + if (pctldesc->complex_pin_desc) > + ret = of_property_count_u32_elems(np, "pins"); > + else > + ret = of_property_count_strings(np, "pins"); ... and has the advantage that you don't have to pass in a complex_pin_desc variable from the driver as the different property name inherently carries this information. 'pins' can then stay a property containing only strings. > if (ret < 0) { > ret = of_property_count_strings(np, "groups"); > if (ret < 0) > @@ -297,11 +305,12 @@ int pinconf_generic_dt_subnode_to_map(struct pinctrl_dev *pctldev, > if (type == PIN_MAP_TYPE_INVALID) > type = PIN_MAP_TYPE_CONFIGS_GROUP; > subnode_target_type = "groups"; > + pins_prop = false; > } else { > if (type == PIN_MAP_TYPE_INVALID) > type = PIN_MAP_TYPE_CONFIGS_PIN; > } > - strings_count = ret; > + items_count = ret; > > ret = of_property_read_string(np, "function", &function); > if (ret < 0) { > @@ -326,17 +335,31 @@ int pinconf_generic_dt_subnode_to_map(struct pinctrl_dev *pctldev, > if (num_configs) > reserve++; > > - reserve *= strings_count; > + reserve *= items_count; > > ret = pinctrl_utils_reserve_map(pctldev, map, reserved_maps, > num_maps, reserve); > if (ret < 0) > goto exit; > > - of_property_for_each_string(np, subnode_target_type, prop, group) { > + items_name = kmalloc_array(items_count, sizeof(char *), GFP_KERNEL); > + if (!items_name) > + goto exit; > + if (pctldesc->complex_pin_desc && pins_prop) { > + of_property_for_each_u32(np, subnode_target_type, prop, cur, val) { > + pin_id = val & PINCTRL_PIN_MASK; > + items_name[i++] = pctldesc->pins[pin_id].name; > + } I don't like that even though pins have numbers here they are converted to strings which the driver later has to search in a list just to convert it back into the number. This is quite inefficient. I guess this could be optimized later, but it would be nice to have the pin number directly in the driver. Sascha -- Pengutronix e.K. | | Industrial Linux Solutions | http://www.pengutronix.de/ | Peiner Str. 6-8, 31137 Hildesheim, Germany | Phone: +49-5121-206917-0 | Amtsgericht Hildesheim, HRA 2686 | Fax: +49-5121-206917-5555 |