On 11/07/18 11:12, Aapo Vienamo wrote: > On Wed, 11 Jul 2018 10:46:21 +0100 > Jon Hunter <jonathanh@xxxxxxxxxx> wrote: > >> On 11/07/18 10:38, Aapo Vienamo wrote: >>> On Wed, 11 Jul 2018 09:40:01 +0100 >>> Jon Hunter <jonathanh@xxxxxxxxxx> wrote: >>> >>>> On 10/07/18 13:54, Aapo Vienamo wrote: >>>>> Register a pinctrl device and implement get and set functions for >>>>> PIN_CONFIG_LOW_POWER_MODE and PIN_CONFIG_POWER_SOURCE parameters. >>>>> >>>>> Signed-off-by: Aapo Vienamo <avienamo@xxxxxxxxxx> >>>>> --- >>>>> drivers/soc/tegra/pmc.c | 192 +++++++++++++++++++++++++++++++++++++++++++++++- >>>>> 1 file changed, 190 insertions(+), 2 deletions(-) >>>>> >>>>> diff --git a/drivers/soc/tegra/pmc.c b/drivers/soc/tegra/pmc.c >>>>> index b833334..8677391 100644 >>>>> --- a/drivers/soc/tegra/pmc.c >>>>> +++ b/drivers/soc/tegra/pmc.c >>>>> @@ -33,6 +33,9 @@ >>>>> #include <linux/of_address.h> >>>>> #include <linux/of_clk.h> >>>>> #include <linux/of_platform.h> >>>>> +#include <linux/pinctrl/pinctrl.h> >>>>> +#include <linux/pinctrl/pinconf.h> >>>>> +#include <linux/pinctrl/pinconf-generic.h> >>>>> #include <linux/platform_device.h> >>>>> #include <linux/pm_domain.h> >>>>> #include <linux/reboot.h> >>>>> @@ -45,6 +48,8 @@ >>>>> #include <soc/tegra/fuse.h> >>>>> #include <soc/tegra/pmc.h> >>>>> >>>>> +#include <dt-bindings/pinctrl/pinctrl-tegra-io-pad.h> >>>>> + >>>>> #define PMC_CNTRL 0x0 >>>>> #define PMC_CNTRL_INTR_POLARITY BIT(17) /* inverts INTR polarity */ >>>>> #define PMC_CNTRL_CPU_PWRREQ_OE BIT(16) /* CPU pwr req enable */ >>>>> @@ -162,6 +167,9 @@ struct tegra_pmc_soc { >>>>> const struct tegra_io_pad_soc *io_pads; >>>>> unsigned int num_io_pads; >>>>> >>>>> + const struct pinctrl_pin_desc *pin_descs; >>>>> + unsigned int num_pin_descs; >>>>> + >>>>> const struct tegra_pmc_regs *regs; >>>>> void (*init)(struct tegra_pmc *pmc); >>>>> void (*setup_irq_polarity)(struct tegra_pmc *pmc, >>>>> @@ -220,6 +228,8 @@ struct tegra_pmc { >>>>> DECLARE_BITMAP(powergates_available, TEGRA_POWERGATE_MAX); >>>>> >>>>> struct mutex powergates_lock; >>>>> + >>>>> + struct pinctrl_dev *pctl_dev; >>>>> }; >>>>> >>>>> static struct tegra_pmc *pmc = &(struct tegra_pmc) { >>>>> @@ -1400,6 +1410,145 @@ static void tegra_pmc_init_tsense_reset(struct tegra_pmc *pmc) >>>>> of_node_put(np); >>>>> } >>>>> >>>>> +static int tegra_io_pad_pinctrl_get_groups_count(struct pinctrl_dev *pctl_dev) >>>>> +{ >>>>> + return pmc->soc->num_io_pads; >>>>> +} >>>>> + >>>>> +static const char *tegra_io_pad_pinctrl_get_group_name( >>>>> + struct pinctrl_dev *pctl, unsigned int group) >>>>> +{ >>>>> + return pmc->soc->io_pads[group].name; >>>>> +} >>>>> + >>>>> +static int tegra_io_pad_pinctrl_get_group_pins(struct pinctrl_dev *pctl_dev, >>>>> + unsigned int group, >>>>> + const unsigned int **pins, >>>>> + unsigned int *num_pins) >>>>> +{ >>>>> + *pins = &pmc->soc->io_pads[group].id; >>>>> + *num_pins = 1; >>>>> + return 0; >>>>> +} >>>>> + >>>>> +static const struct pinctrl_ops tegra_io_pad_pinctrl_ops = { >>>>> + .get_groups_count = tegra_io_pad_pinctrl_get_groups_count, >>>>> + .get_group_name = tegra_io_pad_pinctrl_get_group_name, >>>>> + .get_group_pins = tegra_io_pad_pinctrl_get_group_pins, >>>>> + .dt_node_to_map = pinconf_generic_dt_node_to_map_pin, >>>>> + .dt_free_map = pinconf_generic_dt_free_map, >>>>> +}; >>>>> + >>>>> +static int tegra_io_pad_pinconf_get(struct pinctrl_dev *pctl_dev, >>>>> + unsigned int pin, unsigned long *config) >>>>> +{ >>>>> + const struct tegra_io_pad_soc *pad = tegra_io_pad_find(pmc, pin); >>>>> + enum pin_config_param param = pinconf_to_config_param(*config); >>>>> + int ret; >>>>> + u32 arg; >>>>> + >>>>> + switch (param) { >>>>> + case PIN_CONFIG_POWER_SOURCE: >>>>> + ret = tegra_io_pad_get_voltage(pad->id); >>>>> + if (ret == TEGRA_IO_PAD_1800000UV) >>>>> + arg = TEGRA_IO_PAD_VOLTAGE_1V8; >>>>> + else if (ret == TEGRA_IO_PAD_3300000UV) >>>>> + arg = TEGRA_IO_PAD_VOLTAGE_3V3; >>>> >>>> It looks like we have two definitions for the same thing here. >>>> Can we get rid of one of these? >>> >>> They are indeed pretty much the same thing, however >>> TEGRA_IO_PAD_1800000UV is part of the pmc.h api and >>> TEGRA_IO_PAD_VOLTAGE_1V8 comes from the device tree binding abi headers. >>> The key difference between them is that TEGRA_IO_PAD_VOLTAGE_1V8 is a >>> macro and TEGRA_IO_PAD_1800000UV is an enum and it's used like this in >>> the pmc api: >>> >>> int tegra_io_pad_set_voltage(enum tegra_io_pad id, >>> enum tegra_io_pad_voltage voltage); >>> >>> where it's obvious that the function takes this enum as an argument. If >>> the enum definitions were replaced with macros usable from the dts >>> files, the enum from the function prototype would have to be changed to >>> an int. This makes the api more ambiguous because other parts of the >>> kernel have the convention where "int voltage" means voltage in >>> microvolts. >> >> Raises the question, do we need these legacy APIs? I don't see them being >> used. >> > > True, however other parts of the pmc pad configuration API are used, for > example tegra_io_pad_power_enable() and _disable() are used by the sor > driver. > > I see couple of options here: For sake of consistency we could merge > this as is and later replace all of the existing usage of the pmc.h pad > control API with pinctrl in a separate series and clean that part up. > Or I could remove the pad voltage configuration API as part of this > series and leave the pmc.h pad power state API the way it is. Or I could > transition the existing usage of pmc pad control to use the pinctrl API > but I'm not sure if that's within the scope of this series. > > Any thoughts? My preference would be option #2 ... "remove the pad voltage configuration API as part of this series and leave the pmc.h pad power state API the way it is". If someone is using the tegra_io_pad_power_enable/disable then we cannot get rid of it, but we can get rid of the other legacy APIs. Cheers Jon -- nvpublic -- To unsubscribe from this list: send the line "unsubscribe devicetree" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html