On Sat, Nov 17, 2018 at 09:28:37AM +0100, Loic Poulain wrote: > The PM816 module is a versatile PMIC with many diverse functions > integrated, including, a watchdog. > This watchdog is subcomponent of the PON (Power On) peripheral, > in the same way as pwrkey/resin buttons. > It works with two timers (2-stages), the first one generates an > IRQ to the main SoC (APQ8016/MSM8916), the second one performs > the reset. > > This driver expects the following device hierachy: > [pm8916]->[pm8916-pon]->[pm8916-wdt] > > It uses the pm8916 regmap to access PM8916 registers. > > Signed-off-by: Loic Poulain <loic.poulain@xxxxxxxxxx> > --- > drivers/watchdog/Kconfig | 8 ++ > drivers/watchdog/Makefile | 1 + > drivers/watchdog/pm8916_wdt.c | 177 ++++++++++++++++++++++++++++++++++++++++++ > 3 files changed, 186 insertions(+) > create mode 100644 drivers/watchdog/pm8916_wdt.c > > diff --git a/drivers/watchdog/Kconfig b/drivers/watchdog/Kconfig > index 2d64333..66bab61 100644 > --- a/drivers/watchdog/Kconfig > +++ b/drivers/watchdog/Kconfig > @@ -847,6 +847,14 @@ config SPRD_WATCHDOG > Say Y here to include watchdog timer supported > by Spreadtrum system. > > +config PM8916_WATCHDOG > + tristate "QCOM pm8916 pmic watchdog" > + depends on OF && MFD_SPMI_PMIC > + select WATCHDOG_CORE > + help > + Say Y here to include support watchdog timer embedded into the > + pm8916 module. > + > # X86 (i386 + ia64 + x86_64) Architecture > > config ACQUIRE_WDT > diff --git a/drivers/watchdog/Makefile b/drivers/watchdog/Makefile > index f69cdff..cc90e72 100644 > --- a/drivers/watchdog/Makefile > +++ b/drivers/watchdog/Makefile > @@ -92,6 +92,7 @@ obj-$(CONFIG_STM32_WATCHDOG) += stm32_iwdg.o > obj-$(CONFIG_UNIPHIER_WATCHDOG) += uniphier_wdt.o > obj-$(CONFIG_RTD119X_WATCHDOG) += rtd119x_wdt.o > obj-$(CONFIG_SPRD_WATCHDOG) += sprd_wdt.o > +obj-$(CONFIG_PM8916_WATCHDOG) += pm8916_wdt.o > > # X86 (i386 + ia64 + x86_64) Architecture > obj-$(CONFIG_ACQUIRE_WDT) += acquirewdt.o > diff --git a/drivers/watchdog/pm8916_wdt.c b/drivers/watchdog/pm8916_wdt.c > new file mode 100644 > index 0000000..36f9763 > --- /dev/null > +++ b/drivers/watchdog/pm8916_wdt.c > @@ -0,0 +1,177 @@ > +/* > + * Watchdog driver for QCOM PM8916 > + * > + * Copyright (C) 2018 Loic Poulain, Linaro <loic.poulain@xxxxxxxxxx> > + * > + * This program is free software; you can redistribute it and/or modify > + * it under the terms of the GNU General Public License version 2 and > + * only version 2 as published by the Free Software Foundation. > + * > + * 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. Please use the SPDX license identifier. > + */ > + > +#include <linux/bitops.h> > +#include <linux/kernel.h> > +#include <linux/module.h> > +#include <linux/of.h> > +#include <linux/property.h> > +#include <linux/platform_device.h> > +#include <linux/watchdog.h> > +#include <linux/regmap.h> > + Aplhabetic order, please. > +#define PM8916_WDT_DEFAULT_TIMEOUT 32 > + > +#define PON_PMIC_WD_RESET_S1_TIMER 0x54 > +#define PON_PMIC_WD_RESET_S2_TIMER 0x55 > +#define PON_PMIC_WD_RESET_S2_CTL 0x56 > +#define PON_PMIC_WD_RESET_S2_CTL2 0x57 > +#define PON_PMIC_WD_RESET_PET 0x58 > + > +#define S2_RESET_EN_MASK BIT(7) > +#define S2_RESET_EN (1 << 7) Please use tabs. Also, why (1 << 7) instead of BIT(7) ? Overall, since this is a single bit, having both EN_MASK and EN seems unnecessary, but if you do, > + > +#define WATCHDOG_PET_MASK BIT(0) > + please do the same here. > +struct pm8916_wdt { > + struct device *dev; > + struct regmap *regmap; > + struct watchdog_device wdev; > + u32 baseaddr; > +}; > + > +static int pm8916_wdt_start(struct watchdog_device *wdev) > +{ > + struct pm8916_wdt *wdt = watchdog_get_drvdata(wdev); > + > + return regmap_update_bits(wdt->regmap, > + wdt->baseaddr + PON_PMIC_WD_RESET_S2_CTL2, > + S2_RESET_EN_MASK, S2_RESET_EN); > +} > + > +static int pm8916_wdt_stop(struct watchdog_device *wdev) > +{ > + struct pm8916_wdt *wdt = watchdog_get_drvdata(wdev); > + > + return regmap_update_bits(wdt->regmap, > + wdt->baseaddr + PON_PMIC_WD_RESET_S2_CTL2, > + S2_RESET_EN_MASK, 0); > +} > + > +static int pm8916_wdt_ping(struct watchdog_device *wdev) > +{ > + struct pm8916_wdt *wdt = watchdog_get_drvdata(wdev); > + > + return regmap_update_bits(wdt->regmap, > + wdt->baseaddr + PON_PMIC_WD_RESET_PET, > + WATCHDOG_PET_MASK, 1); The "1" is really supposed to be BIT(0). > +} > + > +static int pm8916_wdt_set_timeout(struct watchdog_device *wdev, > + unsigned int timeout) > +{ > + struct pm8916_wdt *wdt = watchdog_get_drvdata(wdev); > + int err; > + > + wdev->timeout = timeout; > + > + /* This is a two stages watchdog, set stage-1 timer generate an irq and Please use standard multi-line comments. > + * stage-2 trigger the reset. In order to respect expected timeout, set > + * stage-1 timer to timeout and stage-2 to 0. > + */ > + err = regmap_write(wdt->regmap, > + wdt->baseaddr + PON_PMIC_WD_RESET_S1_TIMER, > + timeout); > + if (err) > + return err; > + > + return regmap_write(wdt->regmap, > + wdt->baseaddr + PON_PMIC_WD_RESET_S2_TIMER, > + 0); The watchdog subsystem does support pretimeout, which seems to match the description here. Not mandatory to support it, but it might make sense. > +} > + > +static const struct watchdog_info pm8916_wdt_ident = { > + .options = WDIOF_SETTIMEOUT | WDIOF_KEEPALIVEPING, No WDIOF_MAGICCLOSE ? > + .identity = "QCOM PM8916 PON WDT", > +}; > + > +static const struct watchdog_ops pm8916_wdt_ops = { > + .owner = THIS_MODULE, > + .start = pm8916_wdt_start, > + .stop = pm8916_wdt_stop, > + .ping = pm8916_wdt_ping, > + .set_timeout = pm8916_wdt_set_timeout, > +}; > + > +static int pm8916_wdt_probe(struct platform_device *pdev) > +{ > + struct pm8916_wdt *wdt; > + struct device *parent; > + int err; > + > + wdt = devm_kzalloc(&pdev->dev, sizeof(*wdt), GFP_KERNEL); > + if (!wdt) > + return -ENOMEM; > + > + wdt->dev = &pdev->dev; > + parent = pdev->dev.parent; > + > + /* > + * The pm8916-pon-wdt is a child of the pon device, which is a child > + * of the pm8916 mfd device. We want access to the pm8916 registers. > + * Retrieve regmap from pm8916 (parent->parent) and base address > + * from pm8916-pon (pon). > + */ > + wdt->regmap = dev_get_regmap(parent->parent, NULL); > + if (!wdt->regmap) { > + dev_err(&pdev->dev, "failed to locate regmap\n"); > + return -ENODEV; > + } > + > + err = device_property_read_u32(parent, "reg", &wdt->baseaddr); > + if (err) { > + dev_err(&pdev->dev, "failed to get address\n"); ... and the user will wonder "what address" > + return err; > + } > + > + /* Configure watchdog to hard-reset mode */ > + err = regmap_write(wdt->regmap, > + wdt->baseaddr + PON_PMIC_WD_RESET_S2_CTL, 0x07); This is where symbolic defines come in handy, even if used only once. > + if (err) { > + dev_err(&pdev->dev, "failed configure watchdog\n"); > + return err; > + } > + > + wdt->wdev.info = &pm8916_wdt_ident, > + wdt->wdev.ops = &pm8916_wdt_ops, > + wdt->wdev.parent = &pdev->dev; > + wdt->wdev.min_timeout = 1; > + wdt->wdev.max_timeout = 127; Is this the real range supported by the hardware ? > + wdt->wdev.timeout = PM8916_WDT_DEFAULT_TIMEOUT; > + watchdog_set_drvdata(&wdt->wdev, wdt); > + > + pm8916_wdt_set_timeout(&wdt->wdev, wdt->wdev.timeout); > + Is timeout initialization via fdt not supported on purpose ? > + return watchdog_register_device(&wdt->wdev); devm_watchdog_register_device(), or driver removal may result in an unpleasant surprise. > +} > + > +static const struct of_device_id pm8916_wdt_id_table[] = { > + { .compatible = "qcom,pm8916-wdt" }, > + { } > +}; > +MODULE_DEVICE_TABLE(of, pm8916_wdt_id_table); > + > +static struct platform_driver pm8916_wdt_driver = { > + .probe = pm8916_wdt_probe, > + .driver = { > + .name = "pm8916-wdt", > + .of_match_table = of_match_ptr(pm8916_wdt_id_table), > + }, > +}; > +module_platform_driver(pm8916_wdt_driver); > + > +MODULE_AUTHOR("Loic Poulain <loic.poulain@xxxxxxxxxx>"); > +MODULE_DESCRIPTION("Qualcomm pm8916 watchdog driver"); > +MODULE_LICENSE("GPL v2"); > -- > 2.7.4 >