From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1752409AbaGKJHa (ORCPT ); Fri, 11 Jul 2014 05:07:30 -0400 Received: from mail-ie0-f180.google.com ([209.85.223.180]:49457 "EHLO mail-ie0-f180.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752247AbaGKJHY (ORCPT ); Fri, 11 Jul 2014 05:07:24 -0400 Date: Fri, 11 Jul 2014 10:07:18 +0100 From: Lee Jones To: Stanimir Varbanov Cc: linux-arm-msm@vger.kernel.org, linux-kernel@vger.kernel.org, devicetree@vger.kernel.org, Rob Herring , Kumar Gala , Grant Likely , Courtney Cavin , Josh Cartwright Subject: Re: [RFC PATCH v2 1/5] mfd: qpnp: add support for Qualcomm QPNP PMICs Message-ID: <20140711090718.GA30519@lee--X1> References: <1404393243-7324-1-git-send-email-svarbanov@mm-sol.com> <1404393243-7324-2-git-send-email-svarbanov@mm-sol.com> <20140709143426.GC2635@lee--X1> <53BD5EC5.3080801@mm-sol.com> <20140710083654.GQ2635@lee--X1> <53BEB1DB.6010301@mm-sol.com> MIME-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Disposition: inline Content-Transfer-Encoding: 8bit In-Reply-To: <53BEB1DB.6010301@mm-sol.com> User-Agent: Mutt/1.5.21 (2010-09-15) Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On Thu, 10 Jul 2014, Stanimir Varbanov wrote: > >>>> The Qualcomm QPNP PMIC chips are components used with the > >>>> Snapdragon 800 series SoC family. This driver exists > >>>> largely as a glue mfd component, it exists to be an owner > >>>> of an SPMI regmap for children devices described in > >>>> device tree. > >>>> > >>>> Signed-off-by: Josh Cartwright > >>>> Signed-off-by: Stanimir Varbanov > >>>> --- > >>>> drivers/mfd/Kconfig | 15 ++++++ > >>>> drivers/mfd/Makefile | 1 + > >>>> drivers/mfd/qpnp-spmi.c | 129 +++++++++++++++++++++++++++++++++++++++++++++++ > >>>> 3 files changed, 145 insertions(+), 0 deletions(-) > >>>> create mode 100644 drivers/mfd/qpnp-spmi.c > >>>> > >>>> diff --git a/drivers/mfd/Kconfig b/drivers/mfd/Kconfig > >>>> index ee8204c..258b733 100644 > >>>> --- a/drivers/mfd/Kconfig > >>>> +++ b/drivers/mfd/Kconfig > >>>> @@ -524,6 +524,21 @@ config MFD_PM8921_CORE > >>>> Say M here if you want to include support for PM8921 chip as a module. > >>>> This will build a module called "pm8921-core". > >>>> > >>>> +config MFD_QPNP_SPMI > >>>> + tristate "Qualcomm QPNP SPMI PMIC" > >>>> + depends on ARCH_QCOM || COMPILE_TEST > >>>> + depends on OF > >>>> + select MFD_CORE > >>>> + select REGMAP_SPMI > >>>> + help > >>>> + This enables support for the Qualcomm QPNP SPMI PMICs. > >>>> + These PMICs are currently used with the Snapdragon 800 series of > >>>> + SoCs. Note, that this will only be useful paired with descriptions > >>>> + of the independent functions as children nodes in the device tree. > >>> > > > > >>> > >>>> + * This program is distributed in the hope that it will be useful, > >>>> + * but WITHOUT ANY WARRANTY; without even the implied warranty of > >>>> + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the > >>>> + * GNU General Public License for more details. > >>>> + */ > >>>> +#include > >>>> +#include > >>>> +#include > >>>> +#include > >>>> +#include > >>>> +#include > >>>> +#include > >>>> + > >>>> +#define QPNP_RESOURCE_SIZE 256 > >>>> + > >>>> +static const struct regmap_config qpnp_regmap_config = { > >>>> + .reg_bits = 16, > >>>> + .val_bits = 8, > >>>> + .max_register = 0xffff, > >>>> +}; > >>>> + > >>>> +static int qpnp_index_to_resource(struct device_node *np, int index, > >>>> + struct resource *res) > >>>> +{ > >>>> + const char *name = NULL; > >>>> + const __be32 *addrp; > >>>> + u64 addr; > >>>> + > >>>> + addrp = of_get_address(np, index, NULL, NULL); > >>>> + if (!addrp) > >>>> + return -EINVAL; > >>>> + > >>>> + addr = of_read_number(addrp, 1); > >>>> + if (addr == OF_BAD_ADDR) > >>>> + return -EINVAL; > >>>> + > >>>> + of_property_read_string_index(np, "reg-names", index, &name); > >>>> + > >>>> + res->start = addr; > >>>> + res->end = addr + QPNP_RESOURCE_SIZE - 1; > >>>> + res->flags = IORESOURCE_REG; > >>>> + res->name = name ? name : np->name; > >>>> + > >>>> + return 0; > >>>> +} > >>>> + > >>>> +static int qpnp_add_device(struct spmi_device *root, struct device_node *child) > >>>> +{ > >>>> + struct mfd_cell cell = {}; > >>>> + struct resource *res, *r; > >>>> + int num_resources = 0; > >>>> + const char *compat; > >>>> + int ret, i; > >>>> + > >>>> + compat = of_get_property(child, "compatible", NULL); > >>>> + if (!compat) > >>>> + return -ENODEV; > >>>> + > >>>> + while (of_get_address(child, num_resources, NULL, NULL)) > >>>> + num_resources++; > >>>> + > >>>> + if (!num_resources) > >>>> + return -ENODEV; > >>>> + > >>>> + res = kcalloc(num_resources, sizeof(*res), GFP_KERNEL); > >>>> + if (!res) > >>>> + return -ENOMEM; > >>>> + > >>>> + r = res; > >>>> + for (i = 0; i < num_resources; i++, r++) > >>>> + qpnp_index_to_resource(child, i, r); > >>>> + > >>>> + cell.name = kasprintf(GFP_KERNEL, "%x.%04x.%s", root->usid, > >>>> + (u16)res[0].start, child->name); > >>>> + cell.of_compatible = compat; > >>>> + cell.num_resources = num_resources; > >>>> + cell.resources = res; > >>>> + > >>>> + ret = mfd_add_devices(&root->dev, PLATFORM_DEVID_NONE, &cell, 1, > >>>> + NULL, 0, NULL); > >>>> + > >>>> + kfree(res); > >>>> + kfree(cell.name); > >>>> + > >>>> + return ret; > >>>> +} > >>>> + > >>>> +static int qpnp_probe(struct spmi_device *sdev) > >>>> +{ > >>>> + struct device_node *root = sdev->dev.of_node; > >>>> + struct device_node *child; > >>>> + struct regmap *regmap; > >>>> + > >>>> + regmap = devm_regmap_init_spmi_ext(sdev, &qpnp_regmap_config); > >>>> + if (IS_ERR(regmap)) > >>>> + return PTR_ERR(regmap); > >>>> + > >>>> + for_each_available_child_of_node(root, child) > >>>> + qpnp_add_device(sdev, child); > >>> > >>> This entire driver looks like a re-write of of_platform_populate(). > >>> > >>> Why? > >> > >> of_platform_populate is not used because the PMIC function resources are > >> non-translatable. You can see that the resources are of type > >> IORESOURCE_REG (qpnp_index_to_resource()) not IORESOURCE_MEM or _IO. > >> > >> The whole point of this mfd driver is to parse devicetree to prepare > >> resources for every child and create platform device for it through > >> mfd_add_devices(). Then the PMIC function driver got its resources and > >> use them as register addresses passed to regmap. These register accesses > >> hits the SPMI controller which is the physical interface between PMIC's > >> and SoC. > > > > I can't help but think that if this is required, it should be part of > > the core OF code, rather than doing your own thing which looks > > frighteningly like existing framework functionality. > > does it make sense to have a common of_mfd code for devicetree parsing? > Is that discussed already? I mean presently the mfd_cell resources and > number of resources are passed by the mfd_add_devices users. Is it > possible to have common code which parses devicetree sub-nodes of the > parent device_node and fill resources/create platform devices for them. I'm not sure it does. Normally users _either_ represent devices in MFD cells from within the driver _or_ populate using existing DT interfaces i.e. of_platform_populate(). This is the first time I've seen someone attempt to parse the entire MFD node structure from within a driver. After searching for some documentation to try and figure this out, I noticed that you've also missed a patch from your set: mfd: qpnp-spmi: document DT bindings for Qualcomm QPNP PMICs ... which might help enlighten the DT guys. -- Lee Jones Linaro STMicroelectronics Landing Team Lead Linaro.org │ Open source software for ARM SoCs Follow Linaro: Facebook | Twitter | Blog