From: "Opensource [Steve Twiss]" <[email protected]>
Dependencies required for DA9063 HWMON support.
Signed-off-by: Opensource [Steve Twiss] <[email protected]>
Directly set da9063->t_offset.
Let MFD probe succeed even if DA9063_REG_T_OFFSET cannot be read.
Signed-off-by: Vincent Pelletier <[email protected]>
---
Changes in v2:
- registers.h changes moved from patch 2
Originally submitted by Steve Twiss in 2014:
https://marc.info/?l=linux-kernel&m=139560864709852&w=2
drivers/mfd/da9063-core.c | 8 +++++++
include/linux/mfd/da9063/core.h | 3 +++
include/linux/mfd/da9063/registers.h | 34 ++++++++++++++++++++++++++++
3 files changed, 45 insertions(+)
diff --git a/drivers/mfd/da9063-core.c b/drivers/mfd/da9063-core.c
index df407c3afce3..14c2a8df9ae3 100644
--- a/drivers/mfd/da9063-core.c
+++ b/drivers/mfd/da9063-core.c
@@ -197,6 +197,14 @@ int da9063_device_init(struct da9063 *da9063, unsigned int irq)
}
}
+ ret = regmap_read(da9063->regmap, DA9063_REG_T_OFFSET, &da9063->t_offset);
+ if (ret < 0) {
+ da9063->t_offset = 0;
+ dev_warn(da9063->dev,
+ "Temperature trimming value cannot be read (defaulting to 0)\n");
+ ret = 0;
+ }
+
return ret;
}
diff --git a/include/linux/mfd/da9063/core.h b/include/linux/mfd/da9063/core.h
index fa7a43f02f27..c0c57c6f6230 100644
--- a/include/linux/mfd/da9063/core.h
+++ b/include/linux/mfd/da9063/core.h
@@ -85,6 +85,9 @@ struct da9063 {
int chip_irq;
unsigned int irq_base;
struct regmap_irq_chip_data *regmap_irq;
+
+ /* Trimming */
+ int t_offset;
};
int da9063_device_init(struct da9063 *da9063, unsigned int irq);
diff --git a/include/linux/mfd/da9063/registers.h b/include/linux/mfd/da9063/registers.h
index 6e0f66a2e727..297631ddda39 100644
--- a/include/linux/mfd/da9063/registers.h
+++ b/include/linux/mfd/da9063/registers.h
@@ -512,6 +512,7 @@
/* DA9063_REG_GPIO_0_1 (addr=0x15) */
#define DA9063_GPIO0_PIN_MASK 0x03
+#define DA9063_GPIO0_PIN_MASK_SHIFT 0
#define DA9063_GPIO0_PIN_ADCIN1 0x00
#define DA9063_GPIO0_PIN_GPI 0x01
#define DA9063_GPIO0_PIN_GPO_OD 0x02
@@ -523,6 +524,7 @@
#define DA9063_GPIO0_TYPE_GPO_VDD_IO2 0x04
#define DA9063_GPIO0_NO_WAKEUP 0x08
#define DA9063_GPIO1_PIN_MASK 0x30
+#define DA9063_GPIO1_PIN_MASK_SHIFT 4
#define DA9063_GPIO1_PIN_ADCIN2_COMP 0x00
#define DA9063_GPIO1_PIN_GPI 0x10
#define DA9063_GPIO1_PIN_GPO_OD 0x20
@@ -536,6 +538,7 @@
/* DA9063_REG_GPIO_2_3 (addr=0x16) */
#define DA9063_GPIO2_PIN_MASK 0x03
+#define DA9063_GPIO2_PIN_MASK_SHIFT 0
#define DA9063_GPIO2_PIN_ADCIN3 0x00
#define DA9063_GPIO2_PIN_GPI 0x01
#define DA9063_GPIO2_PIN_GPO_PSS 0x02
@@ -851,6 +854,7 @@
#define DA9063_VSYS_VAL_BASE 0x00
/* DA9063_REG_ADC_RES_L (addr=0x37) */
+#define DA9063_ADC_RES_L_SHIFT 6
#define DA9063_ADC_RES_L_BITS 2
#define DA9063_ADC_RES_L_MASK 0xC0
@@ -1014,6 +1018,36 @@
#define DA9063_GPIO_DIM 0x80
#define DA9063_GPIO_PWM_MASK 0x7F
+/* DA9063_REG_ADC_CFG (addr=0xC9) */
+#define DA9063_REG_ADCIN1_CUR_MASK 0x03
+#define DA9063_REG_ADCIN1_CUR_SHIFT 0
+#define DA9063_ADCIN1_CUR_1UA 0x00
+#define DA9063_ADCIN1_CUR_2UA 0x01
+#define DA9063_ADCIN1_CUR_10UA 0x02
+#define DA9063_ADCIN1_CUR_40UA 0x03
+#define DA9063_REG_ADCIN2_CUR_MASK 0x0C
+#define DA9063_REG_ADCIN2_CUR_SHIFT 2
+#define DA9063_ADCIN2_CUR_1UA 0x00
+#define DA9063_ADCIN2_CUR_2UA 0x01
+#define DA9063_ADCIN2_CUR_10UA 0x02
+#define DA9063_ADCIN2_CUR_40UA 0x03
+#define DA9063_REG_ADCIN3_CUR_MASK 0x10
+#define DA9063_REG_ADCIN3_CUR_SHIFT 4
+#define DA9063_ADCIN3_CUR_10UA 0x00
+#define DA9063_ADCIN3_CUR_40UA 0x01
+#define DA9063_REG_ADCIN1_DEB_MASK 0x20
+#define DA9063_REG_ADCIN1_DEB_SHIFT 5
+#define DA9063_ADCIN1_DEB_OFF 0x00
+#define DA9063_ADCIN1_DEB_ON 0x01
+#define DA9063_REG_ADCIN2_DEB_MASK 0x40
+#define DA9063_REG_ADCIN2_DEB_SHIFT 6
+#define DA9063_ADCIN2_DEB_OFF 0x00
+#define DA9063_ADCIN2_DEB_ON 0x01
+#define DA9063_REG_ADCIN3_DEB_MASK 0x80
+#define DA9063_REG_ADCIN3_DEB_SHIFT 7
+#define DA9063_ADCIN3_DEB_OFF 0x00
+#define DA9063_ADCIN3_DEB_ON 0x01
+
/* DA9063_REG_CONFIG_H (addr=0x10D) */
#define DA9063_PWM_CLK_MASK 0x01
#define DA9063_PWM_CLK_PWM2MHZ 0x00
--
2.32.0
From: "Opensource [Steve Twiss]" <[email protected]>
Addition of HWMON documentation for the DA9063 driver.
Signed-off-by: Opensource [Steve Twiss] <[email protected]>
Updated temperature formula, as of datasheet rev 2.3.
Converted to ReStructuredText.
Signed-off-by: Vincent Pelletier <[email protected]>
---
Changes in v2:
- ReST-ified
Originally submitted by Steve Twiss in 2014:
https://marc.info/?l=linux-kernel&m=139560868209856&w=2
Documentation/hwmon/da9063.rst | 73 ++++++++++++++++++++++++++++++++++
1 file changed, 73 insertions(+)
create mode 100644 Documentation/hwmon/da9063.rst
diff --git a/Documentation/hwmon/da9063.rst b/Documentation/hwmon/da9063.rst
new file mode 100644
index 000000000000..aae69c58a5d6
--- /dev/null
+++ b/Documentation/hwmon/da9063.rst
@@ -0,0 +1,73 @@
+Kernel driver da9063-hwmon
+==========================
+
+Supported chips:
+
+ * Dialog Semiconductor DA9063 PMIC
+
+ Prefix: 'da9063'
+
+ Datasheet: Publicly available at the Dialog Semiconductor website:
+
+ http://www.dialog-semiconductor.com/products/power-management/DA9063
+
+Authors:
+ - S Twiss <[email protected]>
+ - Vincent Pelletier <[email protected]>
+
+Description
+-----------
+
+The DA9063 PMIC provides a general purpose ADC with 10 bits of resolution.
+It uses track and hold circuitry with an analogue input multiplexer which
+allows the conversion of up to 9 different inputs:
+
+- Channel 0: VSYS_RES measurement of the system VDD (2.5 - 5.5V)
+- Channel 1: ADCIN1_RES high impedance input (0 - 2.5V)
+- Channel 2: ADCIN2_RES high impedance input (0 - 2.5V)
+- Channel 3: ADCIN3_RES high impedance input (0 - 2.5V)
+- Channel 4: Tjunc measurement of internal temperature sensor
+- Channel 5: VBBAT measurement of the backup battery voltage (0 - 5.0V)
+- Channel 6: N/A Reserved
+- Channel 7: N/A Reserved
+- Channel 8: MON1_RES group 1 internal regulators voltage (0 - 5.0V)
+- Channel 9: MON2_RES group 2 internal regulators voltage (0 - 5.0V)
+- Channel 10: MON3_RES group 3 internal regulators voltage (0 - 5.0V)
+
+The MUX selects from and isolates the 9 inputs and presents the channel to
+be measured to the ADC input. When selected, an input amplifier on the VSYS
+channel subtracts the VDDCORE reference voltage and scales the signal to the
+correct value for the ADC.
+
+The analog ADC includes current sources at ADC_IN1, ADC_IN2 and ADC_IN3 to
+support resistive measurements.
+
+Channels 1, 2 and 3 current source capability can be set through the ADC
+thresholds ADC_CFG register and values for ADCIN1_CUR, ADCIN2_CUR and
+ADCIN3_CUR. Settings for ADCIN1_CUR and ADCIN2_CUR are 1.0, 2.0, 10 and
+40 micro Amps. The setting for ADCIN3_CUR is 10 micro Amps.
+
+Voltage Monitoring
+------------------
+
+The manual measurement allows monitoring of the system voltage VSYS, the
+auxiliary channels ADCIN1, ADCIN2 and ADCIN3, and a VBBAT measurement of
+the backup battery voltage (0 - 5.0V). The manual measurements store 10
+bits of ADC resolution.
+
+The manual ADC measurements attributes described above are supported by
+the driver.
+
+The automatic ADC measurement is not supported by the driver.
+
+Temperature Monitoring
+----------------------
+
+Temperatures are sampled by a 10 bit ADC. Junction temperatures
+are monitored by the ADC channels.
+
+The junction temperature is calculated:
+
+ Degrees celsius = -0.398 * (ADC_RES - T_OFFSET) + 330;
+
+The junction temperature attribute is supported by the driver.
--
2.32.0
From: "Opensource [Steve Twiss]" <[email protected]>
Add the HWMON driver for DA9063
Signed-off-by: Opensource [Steve Twiss] <[email protected]>
Simplify and modernise the code a bit.
Fix logic inversion in detecting conversion end.
Drop support for ADCIN: these are multi-purpose channels and must not
be reconfigured unless explicitly authorised by the board description.
Signed-off-by: Vincent Pelletier <[email protected]>
---
Changes in v2:
- drop of_match_table: this should be meaningless in such sub-function
driver (at least judging by other sub-function drivers for the da9063)
- sort includes
- switch to devm_hwmon_device_register_with_info
- registers.h changes moved to patch 1
- add SPDX header
This patch depends on patch 1/3.
Originally submitted by Steve Twiss in 2014:
https://marc.info/?l=linux-kernel&m=139560868309857&w=2
drivers/hwmon/Kconfig | 10 ++
drivers/hwmon/Makefile | 1 +
drivers/hwmon/da9063-hwmon.c | 275 +++++++++++++++++++++++++++++++++++
3 files changed, 286 insertions(+)
create mode 100644 drivers/hwmon/da9063-hwmon.c
diff --git a/drivers/hwmon/Kconfig b/drivers/hwmon/Kconfig
index 87624902ea80..17244cfaa855 100644
--- a/drivers/hwmon/Kconfig
+++ b/drivers/hwmon/Kconfig
@@ -515,6 +515,16 @@ config SENSORS_DA9055
This driver can also be built as a module. If so, the module
will be called da9055-hwmon.
+config SENSORS_DA9063
+ tristate "Dialog Semiconductor DA9063"
+ depends on MFD_DA9063
+ help
+ If you say yes here you get support for the hardware
+ monitoring features of the DA9063 Power Management IC.
+
+ This driver can also be built as a module. If so, the module
+ will be called da9063-hwmon.
+
config SENSORS_I5K_AMB
tristate "FB-DIMM AMB temperature sensor on Intel 5000 series chipsets"
depends on PCI
diff --git a/drivers/hwmon/Makefile b/drivers/hwmon/Makefile
index 59e78bc212cf..6855711ed9ec 100644
--- a/drivers/hwmon/Makefile
+++ b/drivers/hwmon/Makefile
@@ -60,6 +60,7 @@ obj-$(CONFIG_SENSORS_CORSAIR_CPRO) += corsair-cpro.o
obj-$(CONFIG_SENSORS_CORSAIR_PSU) += corsair-psu.o
obj-$(CONFIG_SENSORS_DA9052_ADC)+= da9052-hwmon.o
obj-$(CONFIG_SENSORS_DA9055)+= da9055-hwmon.o
+obj-$(CONFIG_SENSORS_DA9063) += da9063-hwmon.o
obj-$(CONFIG_SENSORS_DELL_SMM) += dell-smm-hwmon.o
obj-$(CONFIG_SENSORS_DME1737) += dme1737.o
obj-$(CONFIG_SENSORS_DRIVETEMP) += drivetemp.o
diff --git a/drivers/hwmon/da9063-hwmon.c b/drivers/hwmon/da9063-hwmon.c
new file mode 100644
index 000000000000..f020be5d5d6b
--- /dev/null
+++ b/drivers/hwmon/da9063-hwmon.c
@@ -0,0 +1,275 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/* da9063-hwmon.c - Hardware monitor support for DA9063
+ * Copyright (C) 2014 Dialog Semiconductor Ltd.
+ *
+ * This library is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU Library General Public
+ * License as published by the Free Software Foundation; either
+ * version 2 of the License, or (at your option) any later version.
+ *
+ * This library 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
+ * Library General Public License for more details.
+ */
+
+#include <linux/delay.h>
+#include <linux/err.h>
+#include <linux/hwmon.h>
+#include <linux/hwmon-sysfs.h>
+#include <linux/init.h>
+#include <linux/kernel.h>
+#include <linux/mfd/da9063/core.h>
+#include <linux/mod_devicetable.h>
+#include <linux/module.h>
+#include <linux/platform_device.h>
+#include <linux/regmap.h>
+#include <linux/slab.h>
+#include <linux/string.h>
+
+#define DA9063_ADC_RES (1 << (DA9063_ADC_RES_L_BITS + DA9063_ADC_RES_M_BITS))
+#define DA9063_ADC_MAX (DA9063_ADC_RES - 1)
+#define DA9063_2V5 2500
+#define DA9063_5V0 5000
+#define DA9063_5V5 5500
+#define DA9063_TJUNC_M -398
+#define DA9063_TJUNC_O 330000
+#define DA9063_VBBAT_M 2048
+
+enum da9063_adc {
+ DA9063_CHAN_VSYS = DA9063_ADC_MUX_VSYS,
+ DA9063_CHAN_ADCIN1 = DA9063_ADC_MUX_ADCIN1,
+ DA9063_CHAN_ADCIN2 = DA9063_ADC_MUX_ADCIN2,
+ DA9063_CHAN_ADCIN3 = DA9063_ADC_MUX_ADCIN3,
+ DA9063_CHAN_TJUNC = DA9063_ADC_MUX_T_SENSE,
+ DA9063_CHAN_VBBAT = DA9063_ADC_MUX_VBBAT,
+ DA9063_CHAN_LDO_G1 = DA9063_ADC_MUX_LDO_G1,
+ DA9063_CHAN_LDO_G2 = DA9063_ADC_MUX_LDO_G2,
+ DA9063_CHAN_LDO_G3 = DA9063_ADC_MUX_LDO_G3
+};
+
+struct da9063_hwmon {
+ struct da9063 *da9063;
+ struct mutex hwmon_mutex;
+ struct completion adc_ready;
+ signed char tjunc_offset;
+};
+
+static int da9063_adc_manual_read(struct da9063_hwmon *hwmon, int channel)
+{
+ int ret;
+ unsigned char val;
+ unsigned char data[2];
+ int adc_man;
+
+ mutex_lock(&hwmon->hwmon_mutex);
+
+ init_completion(&hwmon->adc_ready);
+
+ val = (channel & DA9063_ADC_MUX_MASK) | DA9063_ADC_MAN;
+ ret = regmap_update_bits(hwmon->da9063->regmap, DA9063_REG_ADC_MAN,
+ DA9063_ADC_MUX_MASK | DA9063_ADC_MAN, val);
+ if (ret < 0)
+ goto err_mread;
+
+ ret = wait_for_completion_timeout(&hwmon->adc_ready,
+ msecs_to_jiffies(1000));
+ if (ret == 0) {
+ ret = -ETIMEDOUT;
+ goto err_mread;
+ }
+
+ ret = regmap_read(hwmon->da9063->regmap, DA9063_REG_ADC_MAN, &adc_man);
+ if (ret < 0)
+ goto err_mread;
+
+ /* data value is not ready */
+ if (adc_man & DA9063_ADC_MAN) {
+ ret = -EINVAL;
+ goto err_mread;
+ }
+
+ ret = regmap_bulk_read(hwmon->da9063->regmap,
+ DA9063_REG_ADC_RES_L, data, 2);
+ if (ret < 0)
+ goto err_mread;
+
+ ret = (data[0] & DA9063_ADC_RES_L_MASK) >> DA9063_ADC_RES_L_SHIFT;
+ ret |= data[1] << DA9063_ADC_RES_L_BITS;
+err_mread:
+ mutex_unlock(&hwmon->hwmon_mutex);
+ return ret;
+}
+
+static irqreturn_t da9063_hwmon_irq_handler(int irq, void *irq_data)
+{
+ struct da9063_hwmon *hwmon = irq_data;
+
+ complete(&hwmon->adc_ready);
+ return IRQ_HANDLED;
+}
+
+static umode_t da9063_is_visible(const void *drvdata, enum
+ hwmon_sensor_types type, u32 attr, int channel)
+{
+ return 0444;
+}
+
+static const enum da9063_adc da9063_in_index[] = {
+ DA9063_CHAN_VSYS, DA9063_CHAN_VBBAT
+};
+
+static const enum da9063_adc da9063_temp_index[] = {
+ DA9063_CHAN_TJUNC
+};
+
+static int da9063_read(struct device *dev, enum hwmon_sensor_types type,
+ u32 attr, int channel, long *val)
+{
+ struct da9063_hwmon *hwmon = dev_get_drvdata(dev);
+ enum da9063_adc adc_channel;
+ int tmp;
+
+ switch (type) {
+ case hwmon_in:
+ if (attr != hwmon_in_input)
+ return -EOPNOTSUPP;
+ adc_channel = da9063_in_index[channel];
+ break;
+ case hwmon_temp:
+ if (attr != hwmon_temp_input)
+ return -EOPNOTSUPP;
+ adc_channel = da9063_temp_index[channel];
+ break;
+ default:
+ return -EOPNOTSUPP;
+ }
+
+ tmp = da9063_adc_manual_read(hwmon, adc_channel);
+ if (tmp < 0)
+ return tmp;
+
+ switch (adc_channel) {
+ case DA9063_CHAN_VSYS:
+ *val = ((DA9063_5V5 - DA9063_2V5) * tmp) / DA9063_ADC_MAX +
+ DA9063_2V5;
+ break;
+ case DA9063_CHAN_TJUNC:
+ tmp -= hwmon->tjunc_offset;
+ *val = DA9063_TJUNC_M * tmp + DA9063_TJUNC_O;
+ break;
+ case DA9063_CHAN_VBBAT:
+ *val = (DA9063_5V0 * tmp) / DA9063_ADC_MAX;
+ break;
+ default:
+ return -EINVAL;
+ }
+
+ return 0;
+}
+
+static const char * const da9063_in_name[] = {
+ "VSYS", "VBBAT"
+};
+
+static const char * const da9063_temp_name[] = {
+ "TJUNC"
+};
+
+static int da9063_read_string(struct device *dev, enum hwmon_sensor_types type,
+ u32 attr, int channel, const char **str)
+{
+ switch (type) {
+ case hwmon_in:
+ if (attr != hwmon_in_label)
+ return -EOPNOTSUPP;
+ *str = da9063_in_name[channel];
+ break;
+ case hwmon_temp:
+ if (attr != hwmon_temp_label)
+ return -EOPNOTSUPP;
+ *str = da9063_temp_name[channel];
+ break;
+ default:
+ return -EOPNOTSUPP;
+ }
+
+ return 0;
+}
+
+static const struct hwmon_ops da9063_ops = {
+ .is_visible = da9063_is_visible,
+ .read = da9063_read,
+ .read_string = da9063_read_string,
+};
+
+static const struct hwmon_channel_info *da9063_channel_info[] = {
+ HWMON_CHANNEL_INFO(chip, HWMON_C_REGISTER_TZ),
+ HWMON_CHANNEL_INFO(in,
+ HWMON_I_INPUT | HWMON_I_LABEL,
+ HWMON_I_INPUT | HWMON_I_LABEL),
+ HWMON_CHANNEL_INFO(temp,
+ HWMON_T_INPUT | HWMON_T_LABEL),
+ NULL
+};
+
+static const struct hwmon_chip_info da9063_chip_info = {
+ .ops = &da9063_ops,
+ .info = da9063_channel_info,
+};
+
+static int da9063_hwmon_probe(struct platform_device *pdev)
+{
+ struct da9063 *da9063 = dev_get_drvdata(pdev->dev.parent);
+ struct da9063_hwmon *hwmon;
+ struct device *hwmon_dev;
+ int irq;
+ int ret;
+
+ hwmon = devm_kzalloc(&pdev->dev, sizeof(struct da9063_hwmon),
+ GFP_KERNEL);
+ if (!hwmon)
+ return -ENOMEM;
+
+ mutex_init(&hwmon->hwmon_mutex);
+ init_completion(&hwmon->adc_ready);
+ hwmon->da9063 = da9063;
+
+ irq = platform_get_irq_byname(pdev, DA9063_DRVNAME_HWMON);
+ if (irq < 0)
+ return irq;
+
+ ret = devm_request_threaded_irq(&pdev->dev, irq, NULL,
+ da9063_hwmon_irq_handler,
+ IRQF_TRIGGER_LOW | IRQF_ONESHOT,
+ "HWMON", hwmon);
+ if (ret) {
+ dev_err(&pdev->dev, "Failed to request IRQ.\n");
+ return ret;
+ }
+
+ platform_set_drvdata(pdev, hwmon);
+
+ /* set trim temperature offset to value read at startup */
+ hwmon->tjunc_offset = (signed char)hwmon->da9063->t_offset;
+
+ hwmon_dev = devm_hwmon_device_register_with_info(&pdev->dev, "da9063",
+ hwmon,
+ &da9063_chip_info,
+ NULL);
+
+ return PTR_ERR_OR_ZERO(hwmon_dev);
+}
+
+static struct platform_driver da9063_hwmon_driver = {
+ .probe = da9063_hwmon_probe,
+ .driver = {
+ .name = DA9063_DRVNAME_HWMON,
+ },
+};
+module_platform_driver(da9063_hwmon_driver);
+
+MODULE_DESCRIPTION("Hardware monitor support device driver for Dialog DA9063");
+MODULE_AUTHOR("S Twiss <[email protected]>");
+MODULE_LICENSE("GPL v2");
+MODULE_ALIAS("platform:" DA9063_DRVNAME_HWMON);
--
2.32.0
On Tue, Jul 06, 2021 at 02:34:47PM +0000, Vincent Pelletier wrote:
> From: "Opensource [Steve Twiss]" <[email protected]>
>
> Dependencies required for DA9063 HWMON support.
>
> Signed-off-by: Opensource [Steve Twiss] <[email protected]>
>
> Directly set da9063->t_offset.
> Let MFD probe succeed even if DA9063_REG_T_OFFSET cannot be read.
>
> Signed-off-by: Vincent Pelletier <[email protected]>
> ---
> Changes in v2:
> - registers.h changes moved from patch 2
Please version your patch series.
>
> Originally submitted by Steve Twiss in 2014:
> https://marc.info/?l=linux-kernel&m=139560864709852&w=2
>
> drivers/mfd/da9063-core.c | 8 +++++++
> include/linux/mfd/da9063/core.h | 3 +++
> include/linux/mfd/da9063/registers.h | 34 ++++++++++++++++++++++++++++
> 3 files changed, 45 insertions(+)
>
> diff --git a/drivers/mfd/da9063-core.c b/drivers/mfd/da9063-core.c
> index df407c3afce3..14c2a8df9ae3 100644
> --- a/drivers/mfd/da9063-core.c
> +++ b/drivers/mfd/da9063-core.c
> @@ -197,6 +197,14 @@ int da9063_device_init(struct da9063 *da9063, unsigned int irq)
> }
> }
>
> + ret = regmap_read(da9063->regmap, DA9063_REG_T_OFFSET, &da9063->t_offset);
> + if (ret < 0) {
> + da9063->t_offset = 0;
> + dev_warn(da9063->dev,
> + "Temperature trimming value cannot be read (defaulting to 0)\n");
> + ret = 0;
> + }
> +
> return ret;
> }
>
> diff --git a/include/linux/mfd/da9063/core.h b/include/linux/mfd/da9063/core.h
> index fa7a43f02f27..c0c57c6f6230 100644
> --- a/include/linux/mfd/da9063/core.h
> +++ b/include/linux/mfd/da9063/core.h
> @@ -85,6 +85,9 @@ struct da9063 {
> int chip_irq;
> unsigned int irq_base;
> struct regmap_irq_chip_data *regmap_irq;
> +
> + /* Trimming */
> + int t_offset;
> };
>
> int da9063_device_init(struct da9063 *da9063, unsigned int irq);
> diff --git a/include/linux/mfd/da9063/registers.h b/include/linux/mfd/da9063/registers.h
> index 6e0f66a2e727..297631ddda39 100644
> --- a/include/linux/mfd/da9063/registers.h
> +++ b/include/linux/mfd/da9063/registers.h
> @@ -512,6 +512,7 @@
>
> /* DA9063_REG_GPIO_0_1 (addr=0x15) */
> #define DA9063_GPIO0_PIN_MASK 0x03
> +#define DA9063_GPIO0_PIN_MASK_SHIFT 0
> #define DA9063_GPIO0_PIN_ADCIN1 0x00
> #define DA9063_GPIO0_PIN_GPI 0x01
> #define DA9063_GPIO0_PIN_GPO_OD 0x02
> @@ -523,6 +524,7 @@
> #define DA9063_GPIO0_TYPE_GPO_VDD_IO2 0x04
> #define DA9063_GPIO0_NO_WAKEUP 0x08
> #define DA9063_GPIO1_PIN_MASK 0x30
> +#define DA9063_GPIO1_PIN_MASK_SHIFT 4
> #define DA9063_GPIO1_PIN_ADCIN2_COMP 0x00
> #define DA9063_GPIO1_PIN_GPI 0x10
> #define DA9063_GPIO1_PIN_GPO_OD 0x20
> @@ -536,6 +538,7 @@
>
> /* DA9063_REG_GPIO_2_3 (addr=0x16) */
> #define DA9063_GPIO2_PIN_MASK 0x03
> +#define DA9063_GPIO2_PIN_MASK_SHIFT 0
> #define DA9063_GPIO2_PIN_ADCIN3 0x00
> #define DA9063_GPIO2_PIN_GPI 0x01
> #define DA9063_GPIO2_PIN_GPO_PSS 0x02
> @@ -851,6 +854,7 @@
> #define DA9063_VSYS_VAL_BASE 0x00
>
> /* DA9063_REG_ADC_RES_L (addr=0x37) */
> +#define DA9063_ADC_RES_L_SHIFT 6
> #define DA9063_ADC_RES_L_BITS 2
> #define DA9063_ADC_RES_L_MASK 0xC0
>
> @@ -1014,6 +1018,36 @@
> #define DA9063_GPIO_DIM 0x80
> #define DA9063_GPIO_PWM_MASK 0x7F
>
> +/* DA9063_REG_ADC_CFG (addr=0xC9) */
> +#define DA9063_REG_ADCIN1_CUR_MASK 0x03
> +#define DA9063_REG_ADCIN1_CUR_SHIFT 0
> +#define DA9063_ADCIN1_CUR_1UA 0x00
> +#define DA9063_ADCIN1_CUR_2UA 0x01
> +#define DA9063_ADCIN1_CUR_10UA 0x02
> +#define DA9063_ADCIN1_CUR_40UA 0x03
> +#define DA9063_REG_ADCIN2_CUR_MASK 0x0C
> +#define DA9063_REG_ADCIN2_CUR_SHIFT 2
> +#define DA9063_ADCIN2_CUR_1UA 0x00
> +#define DA9063_ADCIN2_CUR_2UA 0x01
> +#define DA9063_ADCIN2_CUR_10UA 0x02
> +#define DA9063_ADCIN2_CUR_40UA 0x03
> +#define DA9063_REG_ADCIN3_CUR_MASK 0x10
> +#define DA9063_REG_ADCIN3_CUR_SHIFT 4
> +#define DA9063_ADCIN3_CUR_10UA 0x00
> +#define DA9063_ADCIN3_CUR_40UA 0x01
> +#define DA9063_REG_ADCIN1_DEB_MASK 0x20
> +#define DA9063_REG_ADCIN1_DEB_SHIFT 5
> +#define DA9063_ADCIN1_DEB_OFF 0x00
> +#define DA9063_ADCIN1_DEB_ON 0x01
> +#define DA9063_REG_ADCIN2_DEB_MASK 0x40
> +#define DA9063_REG_ADCIN2_DEB_SHIFT 6
> +#define DA9063_ADCIN2_DEB_OFF 0x00
> +#define DA9063_ADCIN2_DEB_ON 0x01
> +#define DA9063_REG_ADCIN3_DEB_MASK 0x80
> +#define DA9063_REG_ADCIN3_DEB_SHIFT 7
> +#define DA9063_ADCIN3_DEB_OFF 0x00
> +#define DA9063_ADCIN3_DEB_ON 0x01
> +
> /* DA9063_REG_CONFIG_H (addr=0x10D) */
> #define DA9063_PWM_CLK_MASK 0x01
> #define DA9063_PWM_CLK_PWM2MHZ 0x00
> --
> 2.32.0
>
On Tue, Jul 06, 2021 at 02:34:49PM +0000, Vincent Pelletier wrote:
> From: "Opensource [Steve Twiss]" <[email protected]>
>
> Addition of HWMON documentation for the DA9063 driver.
>
> Signed-off-by: Opensource [Steve Twiss] <[email protected]>
>
> Updated temperature formula, as of datasheet rev 2.3.
> Converted to ReStructuredText.
>
> Signed-off-by: Vincent Pelletier <[email protected]>
> ---
> Changes in v2:
> - ReST-ified
>
> Originally submitted by Steve Twiss in 2014:
> https://marc.info/?l=linux-kernel&m=139560868209856&w=2
>
> Documentation/hwmon/da9063.rst | 73 ++++++++++++++++++++++++++++++++++
Needs to be added to index.rst.
> 1 file changed, 73 insertions(+)
> create mode 100644 Documentation/hwmon/da9063.rst
>
> diff --git a/Documentation/hwmon/da9063.rst b/Documentation/hwmon/da9063.rst
> new file mode 100644
> index 000000000000..aae69c58a5d6
> --- /dev/null
> +++ b/Documentation/hwmon/da9063.rst
> @@ -0,0 +1,73 @@
> +Kernel driver da9063-hwmon
> +==========================
> +
> +Supported chips:
> +
> + * Dialog Semiconductor DA9063 PMIC
> +
> + Prefix: 'da9063'
> +
> + Datasheet: Publicly available at the Dialog Semiconductor website:
> +
> + http://www.dialog-semiconductor.com/products/power-management/DA9063
> +
> +Authors:
> + - S Twiss <[email protected]>
> + - Vincent Pelletier <[email protected]>
> +
> +Description
> +-----------
> +
> +The DA9063 PMIC provides a general purpose ADC with 10 bits of resolution.
> +It uses track and hold circuitry with an analogue input multiplexer which
> +allows the conversion of up to 9 different inputs:
> +
> +- Channel 0: VSYS_RES measurement of the system VDD (2.5 - 5.5V)
> +- Channel 1: ADCIN1_RES high impedance input (0 - 2.5V)
> +- Channel 2: ADCIN2_RES high impedance input (0 - 2.5V)
> +- Channel 3: ADCIN3_RES high impedance input (0 - 2.5V)
> +- Channel 4: Tjunc measurement of internal temperature sensor
> +- Channel 5: VBBAT measurement of the backup battery voltage (0 - 5.0V)
> +- Channel 6: N/A Reserved
> +- Channel 7: N/A Reserved
> +- Channel 8: MON1_RES group 1 internal regulators voltage (0 - 5.0V)
> +- Channel 9: MON2_RES group 2 internal regulators voltage (0 - 5.0V)
> +- Channel 10: MON3_RES group 3 internal regulators voltage (0 - 5.0V)
> +
> +The MUX selects from and isolates the 9 inputs and presents the channel to
> +be measured to the ADC input. When selected, an input amplifier on the VSYS
> +channel subtracts the VDDCORE reference voltage and scales the signal to the
> +correct value for the ADC.
> +
> +The analog ADC includes current sources at ADC_IN1, ADC_IN2 and ADC_IN3 to
> +support resistive measurements.
> +
> +Channels 1, 2 and 3 current source capability can be set through the ADC
> +thresholds ADC_CFG register and values for ADCIN1_CUR, ADCIN2_CUR and
> +ADCIN3_CUR. Settings for ADCIN1_CUR and ADCIN2_CUR are 1.0, 2.0, 10 and
> +40 micro Amps. The setting for ADCIN3_CUR is 10 micro Amps.
> +
> +Voltage Monitoring
> +------------------
> +
> +The manual measurement allows monitoring of the system voltage VSYS, the
> +auxiliary channels ADCIN1, ADCIN2 and ADCIN3, and a VBBAT measurement of
> +the backup battery voltage (0 - 5.0V). The manual measurements store 10
> +bits of ADC resolution.
> +
> +The manual ADC measurements attributes described above are supported by
> +the driver.
> +
> +The automatic ADC measurement is not supported by the driver.
> +
> +Temperature Monitoring
> +----------------------
> +
> +Temperatures are sampled by a 10 bit ADC. Junction temperatures
> +are monitored by the ADC channels.
> +
> +The junction temperature is calculated:
> +
> + Degrees celsius = -0.398 * (ADC_RES - T_OFFSET) + 330;
> +
> +The junction temperature attribute is supported by the driver.
> --
> 2.32.0
>
On Tue, Jul 06, 2021 at 02:34:48PM +0000, Vincent Pelletier wrote:
> From: "Opensource [Steve Twiss]" <[email protected]>
>
> Add the HWMON driver for DA9063
>
> Signed-off-by: Opensource [Steve Twiss] <[email protected]>
>
> Simplify and modernise the code a bit.
> Fix logic inversion in detecting conversion end.
> Drop support for ADCIN: these are multi-purpose channels and must not
> be reconfigured unless explicitly authorised by the board description.
>
That is a bit too much information to be after the first Signed-off-by:
tag, and the code changes are too substantial to warrant Steve's
Signed-off-by: without his explicit permission. I'd suggest to drop
the change log and change Steve's Signed-off-by: to something like
Originally-from: or similar.
> Signed-off-by: Vincent Pelletier <[email protected]>
> ---
> Changes in v2:
> - drop of_match_table: this should be meaningless in such sub-function
> driver (at least judging by other sub-function drivers for the da9063)
> - sort includes
> - switch to devm_hwmon_device_register_with_info
> - registers.h changes moved to patch 1
> - add SPDX header
>
> This patch depends on patch 1/3.
FWIW, that is implied by having a patch series.
> Originally submitted by Steve Twiss in 2014:
> https://marc.info/?l=linux-kernel&m=139560868309857&w=2
>
> drivers/hwmon/Kconfig | 10 ++
> drivers/hwmon/Makefile | 1 +
> drivers/hwmon/da9063-hwmon.c | 275 +++++++++++++++++++++++++++++++++++
> 3 files changed, 286 insertions(+)
> create mode 100644 drivers/hwmon/da9063-hwmon.c
>
> diff --git a/drivers/hwmon/Kconfig b/drivers/hwmon/Kconfig
> index 87624902ea80..17244cfaa855 100644
> --- a/drivers/hwmon/Kconfig
> +++ b/drivers/hwmon/Kconfig
> @@ -515,6 +515,16 @@ config SENSORS_DA9055
> This driver can also be built as a module. If so, the module
> will be called da9055-hwmon.
>
> +config SENSORS_DA9063
> + tristate "Dialog Semiconductor DA9063"
> + depends on MFD_DA9063
> + help
> + If you say yes here you get support for the hardware
> + monitoring features of the DA9063 Power Management IC.
> +
> + This driver can also be built as a module. If so, the module
> + will be called da9063-hwmon.
> +
> config SENSORS_I5K_AMB
> tristate "FB-DIMM AMB temperature sensor on Intel 5000 series chipsets"
> depends on PCI
> diff --git a/drivers/hwmon/Makefile b/drivers/hwmon/Makefile
> index 59e78bc212cf..6855711ed9ec 100644
> --- a/drivers/hwmon/Makefile
> +++ b/drivers/hwmon/Makefile
> @@ -60,6 +60,7 @@ obj-$(CONFIG_SENSORS_CORSAIR_CPRO) += corsair-cpro.o
> obj-$(CONFIG_SENSORS_CORSAIR_PSU) += corsair-psu.o
> obj-$(CONFIG_SENSORS_DA9052_ADC)+= da9052-hwmon.o
> obj-$(CONFIG_SENSORS_DA9055)+= da9055-hwmon.o
> +obj-$(CONFIG_SENSORS_DA9063) += da9063-hwmon.o
> obj-$(CONFIG_SENSORS_DELL_SMM) += dell-smm-hwmon.o
> obj-$(CONFIG_SENSORS_DME1737) += dme1737.o
> obj-$(CONFIG_SENSORS_DRIVETEMP) += drivetemp.o
> diff --git a/drivers/hwmon/da9063-hwmon.c b/drivers/hwmon/da9063-hwmon.c
> new file mode 100644
> index 000000000000..f020be5d5d6b
> --- /dev/null
> +++ b/drivers/hwmon/da9063-hwmon.c
> @@ -0,0 +1,275 @@
> +// SPDX-License-Identifier: GPL-2.0-or-later
> +/* da9063-hwmon.c - Hardware monitor support for DA9063
> + * Copyright (C) 2014 Dialog Semiconductor Ltd.
> + *
> + * This library is free software; you can redistribute it and/or
> + * modify it under the terms of the GNU Library General Public
> + * License as published by the Free Software Foundation; either
> + * version 2 of the License, or (at your option) any later version.
> + *
> + * This library 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
> + * Library General Public License for more details.
Drop boilerplate; that is what SPDX is for.
> + */
> +
> +#include <linux/delay.h>
> +#include <linux/err.h>
> +#include <linux/hwmon.h>
> +#include <linux/hwmon-sysfs.h>
> +#include <linux/init.h>
> +#include <linux/kernel.h>
> +#include <linux/mfd/da9063/core.h>
> +#include <linux/mod_devicetable.h>
> +#include <linux/module.h>
> +#include <linux/platform_device.h>
> +#include <linux/regmap.h>
> +#include <linux/slab.h>
> +#include <linux/string.h>
> +
> +#define DA9063_ADC_RES (1 << (DA9063_ADC_RES_L_BITS + DA9063_ADC_RES_M_BITS))
> +#define DA9063_ADC_MAX (DA9063_ADC_RES - 1)
> +#define DA9063_2V5 2500
> +#define DA9063_5V0 5000
> +#define DA9063_5V5 5500
> +#define DA9063_TJUNC_M -398
> +#define DA9063_TJUNC_O 330000
> +#define DA9063_VBBAT_M 2048
> +
> +enum da9063_adc {
> + DA9063_CHAN_VSYS = DA9063_ADC_MUX_VSYS,
> + DA9063_CHAN_ADCIN1 = DA9063_ADC_MUX_ADCIN1,
> + DA9063_CHAN_ADCIN2 = DA9063_ADC_MUX_ADCIN2,
> + DA9063_CHAN_ADCIN3 = DA9063_ADC_MUX_ADCIN3,
> + DA9063_CHAN_TJUNC = DA9063_ADC_MUX_T_SENSE,
> + DA9063_CHAN_VBBAT = DA9063_ADC_MUX_VBBAT,
> + DA9063_CHAN_LDO_G1 = DA9063_ADC_MUX_LDO_G1,
> + DA9063_CHAN_LDO_G2 = DA9063_ADC_MUX_LDO_G2,
> + DA9063_CHAN_LDO_G3 = DA9063_ADC_MUX_LDO_G3
> +};
> +
> +struct da9063_hwmon {
> + struct da9063 *da9063;
> + struct mutex hwmon_mutex;
> + struct completion adc_ready;
> + signed char tjunc_offset;
> +};
> +
> +static int da9063_adc_manual_read(struct da9063_hwmon *hwmon, int channel)
> +{
> + int ret;
> + unsigned char val;
> + unsigned char data[2];
> + int adc_man;
> +
> + mutex_lock(&hwmon->hwmon_mutex);
> +
> + init_completion(&hwmon->adc_ready);
> +
> + val = (channel & DA9063_ADC_MUX_MASK) | DA9063_ADC_MAN;
> + ret = regmap_update_bits(hwmon->da9063->regmap, DA9063_REG_ADC_MAN,
> + DA9063_ADC_MUX_MASK | DA9063_ADC_MAN, val);
> + if (ret < 0)
> + goto err_mread;
> +
> + ret = wait_for_completion_timeout(&hwmon->adc_ready,
> + msecs_to_jiffies(1000));
> + if (ret == 0) {
> + ret = -ETIMEDOUT;
> + goto err_mread;
> + }
> +
> + ret = regmap_read(hwmon->da9063->regmap, DA9063_REG_ADC_MAN, &adc_man);
> + if (ret < 0)
> + goto err_mread;
> +
> + /* data value is not ready */
> + if (adc_man & DA9063_ADC_MAN) {
> + ret = -EINVAL;
-EINVAL seems wrong. Maybe -EIO or -ETIMEDOUT.
> + goto err_mread;
> + }
> +
> + ret = regmap_bulk_read(hwmon->da9063->regmap,
> + DA9063_REG_ADC_RES_L, data, 2);
> + if (ret < 0)
> + goto err_mread;
> +
> + ret = (data[0] & DA9063_ADC_RES_L_MASK) >> DA9063_ADC_RES_L_SHIFT;
> + ret |= data[1] << DA9063_ADC_RES_L_BITS;
> +err_mread:
> + mutex_unlock(&hwmon->hwmon_mutex);
> + return ret;
> +}
> +
> +static irqreturn_t da9063_hwmon_irq_handler(int irq, void *irq_data)
> +{
> + struct da9063_hwmon *hwmon = irq_data;
> +
> + complete(&hwmon->adc_ready);
> + return IRQ_HANDLED;
> +}
> +
> +static umode_t da9063_is_visible(const void *drvdata, enum
> + hwmon_sensor_types type, u32 attr, int channel)
> +{
> + return 0444;
> +}
> +
> +static const enum da9063_adc da9063_in_index[] = {
> + DA9063_CHAN_VSYS, DA9063_CHAN_VBBAT
> +};
> +
> +static const enum da9063_adc da9063_temp_index[] = {
> + DA9063_CHAN_TJUNC
> +};
> +
> +static int da9063_read(struct device *dev, enum hwmon_sensor_types type,
> + u32 attr, int channel, long *val)
> +{
> + struct da9063_hwmon *hwmon = dev_get_drvdata(dev);
> + enum da9063_adc adc_channel;
> + int tmp;
> +
> + switch (type) {
> + case hwmon_in:
> + if (attr != hwmon_in_input)
> + return -EOPNOTSUPP;
> + adc_channel = da9063_in_index[channel];
> + break;
> + case hwmon_temp:
> + if (attr != hwmon_temp_input)
> + return -EOPNOTSUPP;
> + adc_channel = da9063_temp_index[channel];
There is only one channel for temperatures. I don't see value
in reading that value from an array of size 1.
> + break;
> + default:
> + return -EOPNOTSUPP;
> + }
> +
> + tmp = da9063_adc_manual_read(hwmon, adc_channel);
> + if (tmp < 0)
> + return tmp;
> +
> + switch (adc_channel) {
> + case DA9063_CHAN_VSYS:
> + *val = ((DA9063_5V5 - DA9063_2V5) * tmp) / DA9063_ADC_MAX +
> + DA9063_2V5;
> + break;
> + case DA9063_CHAN_TJUNC:
> + tmp -= hwmon->tjunc_offset;
> + *val = DA9063_TJUNC_M * tmp + DA9063_TJUNC_O;
> + break;
> + case DA9063_CHAN_VBBAT:
> + *val = (DA9063_5V0 * tmp) / DA9063_ADC_MAX;
> + break;
> + default:
> + return -EINVAL;
> + }
> +
> + return 0;
> +}
> +
> +static const char * const da9063_in_name[] = {
> + "VSYS", "VBBAT"
> +};
> +
> +static const char * const da9063_temp_name[] = {
> + "TJUNC"
> +};
> +
> +static int da9063_read_string(struct device *dev, enum hwmon_sensor_types type,
> + u32 attr, int channel, const char **str)
> +{
> + switch (type) {
> + case hwmon_in:
> + if (attr != hwmon_in_label)
> + return -EOPNOTSUPP;
> + *str = da9063_in_name[channel];
> + break;
> + case hwmon_temp:
> + if (attr != hwmon_temp_label)
> + return -EOPNOTSUPP;
> + *str = da9063_temp_name[channel];
> + break;
> + default:
> + return -EOPNOTSUPP;
> + }
> +
> + return 0;
> +}
> +
> +static const struct hwmon_ops da9063_ops = {
> + .is_visible = da9063_is_visible,
> + .read = da9063_read,
> + .read_string = da9063_read_string,
> +};
> +
> +static const struct hwmon_channel_info *da9063_channel_info[] = {
> + HWMON_CHANNEL_INFO(chip, HWMON_C_REGISTER_TZ),
> + HWMON_CHANNEL_INFO(in,
> + HWMON_I_INPUT | HWMON_I_LABEL,
> + HWMON_I_INPUT | HWMON_I_LABEL),
> + HWMON_CHANNEL_INFO(temp,
> + HWMON_T_INPUT | HWMON_T_LABEL),
> + NULL
> +};
> +
> +static const struct hwmon_chip_info da9063_chip_info = {
> + .ops = &da9063_ops,
> + .info = da9063_channel_info,
> +};
> +
> +static int da9063_hwmon_probe(struct platform_device *pdev)
> +{
> + struct da9063 *da9063 = dev_get_drvdata(pdev->dev.parent);
> + struct da9063_hwmon *hwmon;
> + struct device *hwmon_dev;
> + int irq;
> + int ret;
> +
> + hwmon = devm_kzalloc(&pdev->dev, sizeof(struct da9063_hwmon),
> + GFP_KERNEL);
> + if (!hwmon)
> + return -ENOMEM;
> +
> + mutex_init(&hwmon->hwmon_mutex);
> + init_completion(&hwmon->adc_ready);
Does this have a useful purpose ? It is initialized again prior to actually
waiting. Or is this to address the potential race with an interrupt firing
prior to hwmon registration ?
> + hwmon->da9063 = da9063;
> +
> + irq = platform_get_irq_byname(pdev, DA9063_DRVNAME_HWMON);
> + if (irq < 0)
> + return irq;
> +
> + ret = devm_request_threaded_irq(&pdev->dev, irq, NULL,
> + da9063_hwmon_irq_handler,
> + IRQF_TRIGGER_LOW | IRQF_ONESHOT,
Is that correct ? The trigger condition is normally provided by
devicetree.
> + "HWMON", hwmon);
> + if (ret) {
> + dev_err(&pdev->dev, "Failed to request IRQ.\n");
> + return ret;
> + }
> +
> + platform_set_drvdata(pdev, hwmon);
Where is this used ?
> +
> + /* set trim temperature offset to value read at startup */
> + hwmon->tjunc_offset = (signed char)hwmon->da9063->t_offset;
Can you explain why this is read in and passed from the mfd driver
and not here ?
> +
> + hwmon_dev = devm_hwmon_device_register_with_info(&pdev->dev, "da9063",
> + hwmon,
> + &da9063_chip_info,
> + NULL);
> +
> + return PTR_ERR_OR_ZERO(hwmon_dev);
> +}
> +
> +static struct platform_driver da9063_hwmon_driver = {
> + .probe = da9063_hwmon_probe,
> + .driver = {
> + .name = DA9063_DRVNAME_HWMON,
> + },
> +};
> +module_platform_driver(da9063_hwmon_driver);
> +
> +MODULE_DESCRIPTION("Hardware monitor support device driver for Dialog DA9063");
> +MODULE_AUTHOR("S Twiss <[email protected]>");
> +MODULE_LICENSE("GPL v2");
> +MODULE_ALIAS("platform:" DA9063_DRVNAME_HWMON);
> --
> 2.32.0
>
Hello,
Thanks a lot for your reviews.
On Tue, 6 Jul 2021 10:42:01 -0700, Guenter Roeck <[email protected]> wrote:
> -EINVAL seems wrong. Maybe -EIO or -ETIMEDOUT.
On this topic, I've been hesitating to change this code to the
following. Would it be acceptable ?
ret = wait_for_completion_timeout(...)
if (ret == 0)
warn[_once](...)
...
if (adc_man & DA9063_ADC_MAN) {
ret = -ETIMEDOUT;
goto err_mread;
}
The warn is to make it easier to debug in case of IRQ issue. The reason
I'm caring is that I happen to have triggered such issue while testing
this driver, as the GPIO and PLIC on the hifive-unmatched seem to
disagree with each other. I debugged this and reported to linux-riscv,
and I believe the issue is not in da9063-hwmon: it also affects
da9063-onkey, and my GPIO-level workaround fixes both.
On a tangential topic: this chip is supposed to complete an ADC cycle
in 10ms, so 1s timeout seems a lot to me. On the one hand it made the
IRQ issue obvious, but on the other hand a safety factor of 100 seems
enormous to me. What would be a usual/reasonable safety factor ? 10 ?
2 ?
> > + ret = devm_request_threaded_irq(&pdev->dev, irq, NULL,
> > + da9063_hwmon_irq_handler,
> > + IRQF_TRIGGER_LOW | IRQF_ONESHOT,
>
> Is that correct ? The trigger condition is normally provided by
> devicetree.
At least it is consistent with the existing and related da9063-onkey:
irq = platform_get_irq_byname(pdev, "ONKEY");
if (irq < 0)
return irq;
error = devm_request_threaded_irq(&pdev->dev, irq,
NULL, da9063_onkey_irq_handler,
IRQF_TRIGGER_LOW | IRQF_ONESHOT,
"ONKEY", onkey);
I am not familiar enough with IRQ handling to tell if IRQF_TRIGGER_LOW
has an actual meaning here: in my understanding the regmap handler
decides how to clear an interrupt based on regmap_irq_chip content, and
this is coming from mfd/da9063-irq.c .
Are both devm_request_threaded_irq() equally wrong ?
> > + /* set trim temperature offset to value read at startup */
> > + hwmon->tjunc_offset = (signed char)hwmon->da9063->t_offset;
>
> Can you explain why this is read in and passed from the mfd driver
> and not here ?
I cannot, at least not with something other than "this is how I found
the code", which I realise is not satisfactory.
I've been holding back on changes as I felt constrained by preserving
the original author's name on the changes (both Author and
Signed-off-by), but this split was indeed bothering me.
Regards,
--
Vincent Pelletier
GPG fingerprint 983A E8B7 3B91 1598 7A92 3845 CAC9 3691 4257 B0C1
On Wed, Jul 07, 2021 at 12:20:45AM +0000, Vincent Pelletier wrote:
> Hello,
>
> Thanks a lot for your reviews.
>
> On Tue, 6 Jul 2021 10:42:01 -0700, Guenter Roeck <[email protected]> wrote:
> > -EINVAL seems wrong. Maybe -EIO or -ETIMEDOUT.
>
> On this topic, I've been hesitating to change this code to the
> following. Would it be acceptable ?
>
> ret = wait_for_completion_timeout(...)
> if (ret == 0)
> warn[_once](...)
> ...
> if (adc_man & DA9063_ADC_MAN) {
> ret = -ETIMEDOUT;
> goto err_mread;
> }
>
> The warn is to make it easier to debug in case of IRQ issue. The reason
"to debug". Then make it a debug message.
> I'm caring is that I happen to have triggered such issue while testing
> this driver, as the GPIO and PLIC on the hifive-unmatched seem to
> disagree with each other. I debugged this and reported to linux-riscv,
> and I believe the issue is not in da9063-hwmon: it also affects
> da9063-onkey, and my GPIO-level workaround fixes both.
>
> On a tangential topic: this chip is supposed to complete an ADC cycle
> in 10ms, so 1s timeout seems a lot to me. On the one hand it made the
> IRQ issue obvious, but on the other hand a safety factor of 100 seems
> enormous to me. What would be a usual/reasonable safety factor ? 10 ?
> 2 ?
Your call, really.
Guenter
>
> > > + ret = devm_request_threaded_irq(&pdev->dev, irq, NULL,
> > > + da9063_hwmon_irq_handler,
> > > + IRQF_TRIGGER_LOW | IRQF_ONESHOT,
> >
> > Is that correct ? The trigger condition is normally provided by
> > devicetree.
>
> At least it is consistent with the existing and related da9063-onkey:
>
> irq = platform_get_irq_byname(pdev, "ONKEY");
> if (irq < 0)
> return irq;
>
> error = devm_request_threaded_irq(&pdev->dev, irq,
> NULL, da9063_onkey_irq_handler,
> IRQF_TRIGGER_LOW | IRQF_ONESHOT,
> "ONKEY", onkey);
>
> I am not familiar enough with IRQ handling to tell if IRQF_TRIGGER_LOW
> has an actual meaning here: in my understanding the regmap handler
> decides how to clear an interrupt based on regmap_irq_chip content, and
> this is coming from mfd/da9063-irq.c .
>
> Are both devm_request_threaded_irq() equally wrong ?
>
No idea. My understanding is that devicetree data is used unless
an explicit method is specified. That is why I was asking _if_ this
is correct, and did not claim that it is wrong.
> > > + /* set trim temperature offset to value read at startup */
> > > + hwmon->tjunc_offset = (signed char)hwmon->da9063->t_offset;
> >
> > Can you explain why this is read in and passed from the mfd driver
> > and not here ?
>
> I cannot, at least not with something other than "this is how I found
> the code", which I realise is not satisfactory.
> I've been holding back on changes as I felt constrained by preserving
> the original author's name on the changes (both Author and
> Signed-off-by), but this split was indeed bothering me.
>
Sorry, that is not a good argument. On the contrary, if I have to assume
that the code has non-technical constraints, I am inclined to just reject
it for that very reason.
Guenter
On Tue, 6 Jul 2021 17:58:00 -0700, Guenter Roeck <[email protected]> wrote:
> On Wed, Jul 07, 2021 at 12:20:45AM +0000, Vincent Pelletier wrote:
> > On Tue, 6 Jul 2021 10:42:01 -0700, Guenter Roeck <[email protected]> wrote:
> > > > + /* set trim temperature offset to value read at startup */
> > > > + hwmon->tjunc_offset = (signed char)hwmon->da9063->t_offset;
> > >
> > > Can you explain why this is read in and passed from the mfd driver
> > > and not here ?
> >
> > I cannot, at least not with something other than "this is how I found
> > the code", which I realise is not satisfactory.
> > I've been holding back on changes as I felt constrained by preserving
> > the original author's name on the changes (both Author and
> > Signed-off-by), but this split was indeed bothering me.
> >
> Sorry, that is not a good argument. On the contrary, if I have to assume
> that the code has non-technical constraints, I am inclined to just reject
> it for that very reason.
These constraints are self-imposed and authorship-motivated: I do not
want to steal credit, and I do not know where the "why did you remove
my name from this ?" threshold is.
So if I am unsure whether a slightly surprising (to my untrained eyes)
but not broken piece of code is actually usual or should be changed, I
erred on the status-quo side in order to make a good-faith effort at
preserving the original author's name on the change.
So in turn, your suggestion to change the Signed-off-by into an
Originally-from removed these constraints.
Is this formulation clearer ?
--
Vincent Pelletier
GPG fingerprint 983A E8B7 3B91 1598 7A92 3845 CAC9 3691 4257 B0C1