On Thu, 1 Jul 2021 13:28:31 +0300 Alexandru Ardelean <ardeleanalex@xxxxxxxxx> wrote: > On Wed, Jun 30, 2021 at 1:07 AM Lad Prabhakar > <prabhakar.mahadev-lad.rj@xxxxxxxxxxxxxx> wrote: > > > > Add ADC driver support for Renesas RZ/G2L A/D converter in SW > > trigger mode. > > > > A/D Converter block is a successive approximation analog-to-digital > > converter with a 12-bit accuracy and supports a maximum of 8 input > > channels. > > Hey, > > Some comments inline. I added a few more on top, but Alex did a good job so it wasn't much! Jonathan > > > > > Signed-off-by: Lad Prabhakar <prabhakar.mahadev-lad.rj@xxxxxxxxxxxxxx> > > Reviewed-by: Biju Das <biju.das.jz@xxxxxxxxxxxxxx> > > --- > > MAINTAINERS | 8 + > > drivers/iio/adc/Kconfig | 10 + > > drivers/iio/adc/Makefile | 1 + > > drivers/iio/adc/rzg2l_adc.c | 489 ++++++++++++++++++++++++++++++++++++ > > 4 files changed, 508 insertions(+) > > create mode 100644 drivers/iio/adc/rzg2l_adc.c > > > > diff --git a/MAINTAINERS b/MAINTAINERS > > index 81e1edeceae4..bee4c3847e01 100644 > > --- a/MAINTAINERS > > +++ b/MAINTAINERS > > @@ -15622,6 +15622,14 @@ L: linux-renesas-soc@xxxxxxxxxxxxxxx > > S: Maintained > > F: drivers/phy/renesas/phy-rcar-gen3-usb*.c > > > > +RENESAS RZ/G2L A/D DRIVER > > +M: Lad Prabhakar <prabhakar.mahadev-lad.rj@xxxxxxxxxxxxxx> > > +L: linux-iio@xxxxxxxxxxxxxxx > > +L: linux-renesas-soc@xxxxxxxxxxxxxxx > > +S: Supported > > +F: Documentation/devicetree/bindings/iio/adc/renesas,rzg2l-adc.yaml > > +F: drivers/iio/adc/rzg2l_adc.c > > + > > RESET CONTROLLER FRAMEWORK > > M: Philipp Zabel <p.zabel@xxxxxxxxxxxxxx> > > S: Maintained > > diff --git a/drivers/iio/adc/Kconfig b/drivers/iio/adc/Kconfig > > index c7946c439612..9408cbf97acc 100644 > > --- a/drivers/iio/adc/Kconfig > > +++ b/drivers/iio/adc/Kconfig > > @@ -887,6 +887,16 @@ config ROCKCHIP_SARADC > > To compile this driver as a module, choose M here: the > > module will be called rockchip_saradc. > > > > +config RZG2L_ADC > > + tristate "Renesas RZ/G2L ADC driver" > > + depends on ARCH_R9A07G044 || COMPILE_TEST > > + help > > + Say yes here to build support for the ADC found in Renesas > > + RZ/G2L family. > > + > > + To compile this driver as a module, choose M here: the > > + module will be called rzg2l_adc. > > + > > config SC27XX_ADC > > tristate "Spreadtrum SC27xx series PMICs ADC" > > depends on MFD_SC27XX_PMIC || COMPILE_TEST > > diff --git a/drivers/iio/adc/Makefile b/drivers/iio/adc/Makefile > > index a226657d19c0..d92bcc9c5fbb 100644 > > --- a/drivers/iio/adc/Makefile > > +++ b/drivers/iio/adc/Makefile > > @@ -82,6 +82,7 @@ obj-$(CONFIG_QCOM_PM8XXX_XOADC) += qcom-pm8xxx-xoadc.o > > obj-$(CONFIG_RCAR_GYRO_ADC) += rcar-gyroadc.o > > obj-$(CONFIG_RN5T618_ADC) += rn5t618-adc.o > > obj-$(CONFIG_ROCKCHIP_SARADC) += rockchip_saradc.o > > +obj-$(CONFIG_RZG2L_ADC) += rzg2l_adc.o > > obj-$(CONFIG_SC27XX_ADC) += sc27xx_adc.o > > obj-$(CONFIG_SPEAR_ADC) += spear_adc.o > > obj-$(CONFIG_STX104) += stx104.o > > diff --git a/drivers/iio/adc/rzg2l_adc.c b/drivers/iio/adc/rzg2l_adc.c > > new file mode 100644 > > index 000000000000..1c58eb8ae1ec > > --- /dev/null > > +++ b/drivers/iio/adc/rzg2l_adc.c > > @@ -0,0 +1,489 @@ > > +// SPDX-License-Identifier: GPL-2.0 > > +/* > > + * RZ/G2L A/D Converter driver > > + * > > + * Copyright (c) 2021 Renesas Electronics Europe GmbH > > + * > > + * Author: Lad Prabhakar <prabhakar.mahadev-lad.rj@xxxxxxxxxxxxxx> > > + */ > > + > > +#include <linux/clk.h> > > +#include <linux/completion.h> > > +#include <linux/delay.h> > > +#include <linux/iio/iio.h> > > +#include <linux/interrupt.h> > > +#include <linux/io.h> > > +#include <linux/module.h> > > +#include <linux/of.h> > > +#include <linux/of_device.h> I'd in general prefer use of generic firmware properties rather than the of variant but I guess you can be fairly sure these devices are always going to be using devicetree, so it doesn't really matter. > > +#include <linux/platform_device.h> > > +#include <linux/reset.h> > > + > > +#define ADM(n) ((n) * 0x4) I'd prefer some sort of driver specific prefix on these defines as some are very generic sounding so it would be good to make it clear they are local to this driver. > > +#define ADM0_ADCE BIT(0) > > +#define ADM0_ADBSY BIT(1) > > +#define ADM0_PWDWNB BIT(2) > > +#define ADM0_SRESB BIT(15) > > +#define ADM1_TRG BIT(0) > > +#define ADM1_MS BIT(2) > > +#define ADM1_BS BIT(4) > > +#define ADM1_EGA_CLEAR ~GENMASK(13, 12) > > +#define ADM2_CHSEL_CLEAR ~GENMASK(7, 0) > > +#define ADM3_ADSMP 0x578 > > +#define ADM3_ADCMP (0xe << 16) > > +#define ADM3_ADIL_CLEAR ~GENMASK(31, 24) > > + > > +#define ADINT 0x20 > > +#define ADINT_CH_CLEAR ~GENMASK(7, 0) > > +#define ADINT_CSEEN BIT(16) > > +#define ADINT_INTS BIT(31) > > +#define ADSTS 0x24 > > +#define ADINT_INTST_MASK GENMASK(7, 0) > > +#define ADSTS_CSEST BIT(16) > > +#define ADIVC 0x28 > > +#define ADIVC_DIVADC_CLEAR ~GENMASK(8, 0) > > +#define ADIVC_DIVADC_4 0x4 > > +#define ADFIL 0x2c > > +#define ADCR(n) (0x30 + ((n) * 0x4)) > > +#define ADCR_AD_MASK GENMASK(11, 0) > > + > > +#define ADC_MAX_CHANNELS 8 > > +#define ADC_CHN_MASK 0x7 > > +#define ADC_TIMEOUT usecs_to_jiffies(1 * 4) > > + > > +enum trigger_mode { > > + SW_TRIGGER = 0, > > + SYNC_TRIGGER, > > + ASYNC_TRIGGER, > > +}; > > this enum could also be removed [for now] given that only SW_TRIGGER > is supported; > > > + > > +struct rzg2l_adc_data { > > + const struct iio_chan_spec *channels; > > + u8 num_channels; > > + u8 trigger; > > +}; > > + > > +struct rzg2l_adc { > > + void __iomem *base; > > + struct clk *pclk; > > + struct clk *adclk; > > + struct reset_control *presetn; > > + struct reset_control *adrstn; > > + struct completion completion; > > + const struct rzg2l_adc_data *data; > > + bool adc_disabled; /* protected with mlock mutex from indio_dev */ > > this adc_disabled flag looks a bit weird; > it seems to guard against this driver being removed to prevent some reads. > technically, this should be protected by IIO core; > so the flag itself (or how it is being used) looks like it doesn't do much; Agreed. If we have paths we are missing, then please let us know. There may well be some as it's been a while since Lars did a bunch of work testing and fixing issues around remove races. > > > + u16 last_val[ADC_MAX_CHANNELS]; > > +}; > > + > > +static unsigned int rzg2l_adc_readl(struct rzg2l_adc *adc, u32 reg) > > +{ > > + return readl(adc->base + reg); Hmm. Bit marginal as to whether these are worthwhile. If you really like them I guess I don't really mind. > > +} > > + > > +static void rzg2l_adc_writel(struct rzg2l_adc *adc, unsigned int reg, u32 val) > > +{ > > + writel(val, adc->base + reg); > > +} > > + > > +static int rzg2l_adc_adclk(struct rzg2l_adc *adc, bool prepare) > > +{ > > + if (prepare) > > + return clk_prepare_enable(adc->adclk); I'd drop this function and call clk_prepare_enable() / clk_disable_unprepare() directly inline. This just makes the code more confusing but implying that the disable_unprepare can fail for example. > > + > > + clk_disable_unprepare(adc->adclk); > > + return 0; > > +} > > + > > +static void rzg2l_adc_pwr(struct rzg2l_adc *adc, bool on) > > +{ > > + u32 reg; > > + > > + reg = rzg2l_adc_readl(adc, ADM(0)); > > + if (on) > > + reg |= ADM0_PWDWNB; > > + else > > + reg &= ~ADM0_PWDWNB; > > + rzg2l_adc_writel(adc, ADM(0), reg); > > + udelay(2); > > +} > > + > > +static void rzg2l_adc_conversion(struct rzg2l_adc *adc, bool start) > > +{ > > + int timeout = 5; > > + u32 reg; > > + > > + /* stop A/D conversion */ > > + reg = rzg2l_adc_readl(adc, ADM(0)); > > + if (start) > > + reg |= ADM0_ADCE; > > + else > > + reg &= ~ADM0_ADCE; > > + rzg2l_adc_writel(adc, ADM(0), reg); > > + > > + if (start) > > + return; > > + > > + do { > > + usleep_range(100, 200); > > + reg = rzg2l_adc_readl(adc, ADM(0)); > > + timeout--; > > + if (!timeout) { > > + pr_err("%s stopping ADC timed out\n", __func__); > > + break; > > + } > > + } while (((reg & ADM0_ADBSY) || (reg & ADM0_ADCE))); > > +} > > + > > +static int rzg2l_adc_read_raw(struct iio_dev *indio_dev, > > + struct iio_chan_spec const *chan, > > + int *val, int *val2, long mask) > > +{ > > + struct rzg2l_adc *adc = iio_priv(indio_dev); > > + u32 reg; > > + int ret; > > + u8 ch; > > + > > + switch (mask) { > > + case IIO_CHAN_INFO_RAW: > > + mutex_lock(&indio_dev->mlock); > > [1] > acquiring indio_dev->mlock directly is discouraged; > this lock is reserved for IIO core logic and will be moved into an > iio_dev_opaque struct eventually; > driver state locks should be defined in struct rzg2l_adc and managed > by the driver to protect it's own internal state; I'd also be tempted to factor this read switch block out to a separate function. That way you can either take the new lock outside of the function (and then the function can do direct returns in error paths) or you can use goto err; and unlock there so that you don't have any risk of forgetting an unlock if this code is refactored in future. > > > > + > > + if (adc->adc_disabled) { > > + mutex_unlock(&indio_dev->mlock); > > + return -EBUSY; > > + } > > + > > + if (rzg2l_adc_readl(adc, ADM(0)) & ADM0_ADBSY) { > > + mutex_unlock(&indio_dev->mlock); > > + return -EBUSY; > > + } > > + > > + ch = chan->channel & ADC_CHN_MASK; > > + /* SW trigger */ > > + reg = rzg2l_adc_readl(adc, ADM(1)); > > + reg &= ADM1_EGA_CLEAR; > > + reg &= ~ADM1_BS; > > + reg |= ADM1_MS; > > + reg &= ~ADM1_TRG; > > + rzg2l_adc_writel(adc, ADM(1), reg); > > + > > + /* select channel */ > > + reg = rzg2l_adc_readl(adc, ADM(2)); > > + reg &= ADM2_CHSEL_CLEAR; > > + reg |= BIT(ch); > > + rzg2l_adc_writel(adc, ADM(2), reg); > > + > > + reg = rzg2l_adc_readl(adc, ADM(3)); > > + reg &= ADM3_ADIL_CLEAR; > > + reg |= ADM3_ADCMP; > > + reg |= ADM3_ADSMP; > > + rzg2l_adc_writel(adc, ADM(3), reg); > > + > > + reg = rzg2l_adc_readl(adc, ADIVC); > > + reg &= ADIVC_DIVADC_CLEAR; > > + reg |= ADIVC_DIVADC_4; > > + rzg2l_adc_writel(adc, ADIVC, reg); > > + > > + reg = rzg2l_adc_readl(adc, ADINT); > > + reg &= ~ADINT_INTS; > > + reg &= ADINT_CH_CLEAR; > > + reg |= ADINT_CSEEN; > > + reg |= BIT(ch); > > + rzg2l_adc_writel(adc, ADINT, reg); > > + > > + rzg2l_adc_pwr(adc, true); > > should all this clock & power management be done in this read function? > it looks like an awful lot just to perform a single read > maybe some PM suspend/resume hooks would be a better idea for these; > > > + > > + ret = rzg2l_adc_adclk(adc, true); > > + if (ret) { > > + rzg2l_adc_pwr(adc, false); > > + mutex_unlock(&indio_dev->mlock); > > + return -EINVAL; > > + } > > + > > + reinit_completion(&adc->completion); > > + > > + rzg2l_adc_conversion(adc, true); > > + > > + if (!wait_for_completion_timeout(&adc->completion, ADC_TIMEOUT)) { > > + reg &= ADINT_CH_CLEAR; > > + rzg2l_adc_writel(adc, ADINT, reg); > > + rzg2l_adc_conversion(adc, false); > > + rzg2l_adc_adclk(adc, false); > > + rzg2l_adc_pwr(adc, false); > > + mutex_unlock(&indio_dev->mlock); > > + return -ETIMEDOUT; > > + } > > + > > + *val = adc->last_val[ch]; > > + rzg2l_adc_conversion(adc, false); > > + rzg2l_adc_adclk(adc, false); > > + rzg2l_adc_pwr(adc, false); > > + mutex_unlock(&indio_dev->mlock); > > + return IIO_VAL_INT; > > + > > + default: > > + return -EINVAL; > > + } > > +} > > + > > +static irqreturn_t rzg2l_adc_isr(int irq, void *dev_id) > > +{ > > + struct rzg2l_adc *adc = (struct rzg2l_adc *)dev_id; No need to cast a void * to anything explicitly. The c spec allows this to always be done via a simple assignment. > > + u8 intst; > > + u32 reg; > > + u8 i; > > + > > + reg = rzg2l_adc_readl(adc, ADSTS); > > + if (reg & ADSTS_CSEST) { > > + rzg2l_adc_writel(adc, ADSTS, reg); > > + return IRQ_HANDLED; > > + } > > + > > + intst = reg & ADINT_INTST_MASK; > > + if (!intst) > > + return IRQ_HANDLED; > > + > > + for (i = 0; i < ADC_MAX_CHANNELS; i++) { > > + if (intst & BIT(i)) > > + adc->last_val[i] = rzg2l_adc_readl(adc, ADCR(i)) & ADCR_AD_MASK; > > + } > > + > > + rzg2l_adc_writel(adc, ADSTS, reg); > > + > > + complete(&adc->completion); > > + > > + return IRQ_HANDLED; > > +} > > + > > +static const struct iio_info rzg2l_adc_iio_info = { > > + .read_raw = rzg2l_adc_read_raw, > > +}; > > + > > +static const char * const rzg2l_adc_channel_name[] = { > > + "adc0", > > + "adc1", > > + "adc2", > > + "adc3", > > + "adc4", > > + "adc5", > > + "adc6", > > + "adc7", > > +}; > > + > > +static int rzg2l_adc_parse_of(struct platform_device *pdev, struct rzg2l_adc *adc) > > +{ > > + struct device_node *node = pdev->dev.of_node; > > + struct iio_chan_spec *chan_array; > > + u8 channels[ADC_MAX_CHANNELS]; > > + struct rzg2l_adc_data *data; > > + int num_channels; > > + int ret; > > + u8 i; > > + > > + data = devm_kzalloc(&pdev->dev, sizeof(*data), GFP_KERNEL); > > + if (!data) > > + return -ENOMEM; > > + > > + num_channels = of_property_count_u8_elems(node, "renesas-rzg2l,adc-channels"); > > + if (num_channels <= 0 || num_channels > ADC_MAX_CHANNELS) > > + return -EINVAL; > > + > > + ret = of_property_read_u8_array(node, "renesas-rzg2l,adc-channels", > > + channels, num_channels); > > + if (ret) > > + return ret; > > + > > + chan_array = devm_kcalloc(&pdev->dev, num_channels, sizeof(*chan_array), > > + GFP_KERNEL); > > + if (!chan_array) > > + return -ENOMEM; > > + > > + for (i = 0; i < num_channels; i++) { > > + chan_array[i].type = IIO_VOLTAGE; > > + chan_array[i].indexed = 1; > > + chan_array[i].channel = channels[i]; > > + chan_array[i].info_mask_separate = BIT(IIO_CHAN_INFO_RAW); > > + chan_array[i].info_mask_shared_by_type = BIT(IIO_CHAN_INFO_SCALE); > > + chan_array[i].datasheet_name = rzg2l_adc_channel_name[i]; > > + } > > + > > + ret = of_property_read_u8(node, "renesas-rzg2l,adc-trigger-mode", > > + &data->trigger); > > + if (ret) > > + data->trigger = SW_TRIGGER; > > + > > + /* we support SW_TRIGGER as of now */ > > + if (data->trigger != SW_TRIGGER) > > + return -EINVAL; > > it would be an idea to remove this data->trigger field and the DT read > for this property and add it when it's supported; > typically these triggers don't get configured via DT; > > > + > > + data->num_channels = num_channels; > > + data->channels = chan_array; > > + adc->data = data; > > + > > + return 0; > > +} > > + > > +static int rzg2l_adc_sw_reset(struct rzg2l_adc *adc) > > +{ > > + int timeout = 5; > > + u32 val; > > + > > + val = rzg2l_adc_readl(adc, ADM(0)); > > + val |= ADM0_SRESB; > > + rzg2l_adc_writel(adc, ADM(0), val); > > + > > + while (!(rzg2l_adc_readl(adc, ADM(0)) & ADM0_SRESB)) { > > + if (!timeout) > > + return -EINVAL; > > maybe -EBUSY is a bit better error code; > > > + timeout--; > > + usleep_range(100, 200); > > + } > > + > > + return 0; > > +} > > + > > +static int rzg2l_adc_probe(struct platform_device *pdev) > > +{ > > + struct iio_dev *indio_dev; > > + struct rzg2l_adc *adc; > > + int ret; > > + int irq; > > + > > + indio_dev = devm_iio_device_alloc(&pdev->dev, sizeof(*adc)); > > + if (!indio_dev) { > > + dev_err(&pdev->dev, "failed allocating iio device\n"); > > this message can be removed; > looks like log spam; > and if it happens, the system will be in a pretty bad state anyway > > > + return -ENOMEM; > > + } > > + > > + adc = iio_priv(indio_dev); > > + if (!adc) > > + return -ENOMEM; > > this check is redundant; > if indio_dev is non-NULL then iio_priv() will be good as well; > > > + > > + ret = rzg2l_adc_parse_of(pdev, adc); > > + if (ret) > > + return -ENOMEM; > > + > > + adc->base = devm_platform_ioremap_resource(pdev, 0); > > + if (IS_ERR(adc->base)) { > > + dev_err(&pdev->dev, "missing mem resource"); > > this message can be removed; > looks like log-spam > > > + return PTR_ERR(adc->base); > > + } > > + > > + irq = platform_get_irq(pdev, 0); > > + if (irq < 0) { > > + dev_err(&pdev->dev, "no irq resource\n"); > > + return irq; > > + } > > + > > + adc->pclk = devm_clk_get(&pdev->dev, "pclk"); > > + if (IS_ERR(adc->pclk)) { > > + dev_err(&pdev->dev, "Failed to get pclk"); > > + return PTR_ERR(adc->pclk); > > + } > > + > > + adc->adclk = devm_clk_get(&pdev->dev, "adclk"); > > + if (IS_ERR(adc->adclk)) { > > + dev_err(&pdev->dev, "Failed to get adclk"); > > + return PTR_ERR(adc->adclk); > > + } > > + > > + adc->adrstn = devm_reset_control_get_exclusive(&pdev->dev, "adrst-n"); > > + if (IS_ERR(adc->adrstn)) { > > + dev_err(&pdev->dev, "failed to get adrstn\n"); > > + return PTR_ERR(adc->adrstn); > > + } > > + > > + adc->presetn = devm_reset_control_get_exclusive(&pdev->dev, "presetn"); > > + if (IS_ERR(adc->presetn)) { > > + dev_err(&pdev->dev, "failed to get presetn\n"); > > + return PTR_ERR(adc->presetn); > > + } > > + > > + ret = reset_control_deassert(adc->adrstn); > > + if (ret) > > + return ret; > > + > > + ret = reset_control_deassert(adc->presetn); > > + if (ret) > > + goto assert_adrstn; > > + > > + ret = clk_prepare_enable(adc->pclk); > > + if (ret) > > + goto assert_presetn; > > + > > + ret = rzg2l_adc_sw_reset(adc); > > + if (ret) > > + goto unprepare_pclk; > > + > > + init_completion(&adc->completion); > > + > > + platform_set_drvdata(pdev, indio_dev); > > + > > + ret = devm_request_irq(&pdev->dev, irq, rzg2l_adc_isr, > > + 0, dev_name(&pdev->dev), adc); > > + if (ret < 0) > > + goto unprepare_pclk; > > + > > + adc->adc_disabled = false; > > + indio_dev->name = dev_name(&pdev->dev); > > indio_dev->name should be the part-name; > since this driver supports a single part, this can be: > > indio_dev->name = "rzg2l-adc"; > > > + indio_dev->dev.parent = &pdev->dev; > > + indio_dev->dev.of_node = pdev->dev.of_node; > > > The 2 assignments above can be removed in the mainline driver. > They should be done in devm_iio_device_alloc() and iio_device_register() > > > > + indio_dev->info = &rzg2l_adc_iio_info; > > + indio_dev->modes = INDIO_DIRECT_MODE; > > + indio_dev->channels = adc->data->channels; > > + indio_dev->num_channels = adc->data->num_channels; > > + > > + ret = iio_device_register(indio_dev); > > + if (ret) > > + goto unprepare_pclk; > > + > > + return 0; > > + > > +unprepare_pclk: > > + clk_disable_unprepare(adc->pclk); > > +assert_presetn: > > + reset_control_assert(adc->presetn); > > +assert_adrstn: > > + reset_control_assert(adc->adrstn); > > + return ret; > > +} > > + > > +static int rzg2l_adc_remove(struct platform_device *pdev) > > +{ > > + struct iio_dev *indio_dev = platform_get_drvdata(pdev); > > + struct rzg2l_adc *adc = iio_priv(indio_dev); > > + > > + mutex_lock(&indio_dev->mlock); > > + adc->adc_disabled = true; > > + mutex_unlock(&indio_dev->mlock); > > + > > + iio_device_unregister(indio_dev); > > + > > + clk_disable_unprepare(adc->pclk); > > + reset_control_assert(adc->presetn); > > + reset_control_assert(adc->adrstn); > > + > > + return 0; > > +} > > + > > +static const struct of_device_id rzg2l_adc_match[] = { > > + { > > + .compatible = "renesas,rzg2l-adc", > > + }, > > + {}, > > comma can be removed; > since this is a null terminator > > > +}; > > +MODULE_DEVICE_TABLE(of, rzg2l_adc_match); > > + > > +static struct platform_driver rzg2l_adc_driver = { > > + .probe = rzg2l_adc_probe, > > + .remove = rzg2l_adc_remove, > > + .driver = { > > + .name = "rzg2l-adc", > > + .of_match_table = rzg2l_adc_match, > > + }, > > +}; > > + > > +module_platform_driver(rzg2l_adc_driver); > > + > > +MODULE_AUTHOR("Lad Prabhakar <prabhakar.mahadev-lad.rj@xxxxxxxxxxxxxx>"); > > +MODULE_DESCRIPTION("Renesas RZ/G2L ADC driver"); > > +MODULE_LICENSE("GPL v2"); > > -- > > 2.17.1 > >