RE: [PATCH 2/2] hwmon: da9052: add support for TSI channel

From: Steve Twiss
Date: Wed Jun 28 2017 - 13:04:01 EST

Next message: Keith Busch: "Re: [PATCH] fs: System memory leak when running HTX with T10 DIF enabled"
Previous message: Paul E. McKenney: "Re: [GIT PULL rcu/next] RCU commits for 4.13"
In reply to: Guenter Roeck: "Re: [PATCH 2/2] hwmon: da9052: add support for TSI channel"
Next in thread: Sebastian Reichel: "[PATCH 1/2] mfd: da9052: fix manual ADC read after timed out read"
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]

Hi Sebastian,

On 26 June 2017 15:49 Sebastian Reichel wrote:
> Subject: [PATCH 2/2] hwmon: da9052: add support for TSI channel
>
> TSI channel has a 4 channel mux connected to it and is normally
> used for touchscreen support. The hardware may alternatively
> use it as general purpose adc.
>
> Signed-off-by: Sebastian Reichel <sebastian.reichel@xxxxxxxxxxxxxxx>
> ---
> .../devicetree/bindings/mfd/da9052-i2c.txt | 8 +
> drivers/hwmon/da9052-hwmon.c | 229
> ++++++++++++++++++++-
> drivers/input/touchscreen/da9052_tsi.c | 5 +
> include/linux/mfd/da9052/da9052.h | 6 +
> 4 files changed, 244 insertions(+), 4 deletions(-)
>
> diff --git a/Documentation/devicetree/bindings/mfd/da9052-i2c.txt
> b/Documentation/devicetree/bindings/mfd/da9052-i2c.txt
> index 9554292dc6cb..94f3bb1d9a0a 100644
> --- a/Documentation/devicetree/bindings/mfd/da9052-i2c.txt
> +++ b/Documentation/devicetree/bindings/mfd/da9052-i2c.txt
> @@ -4,6 +4,14 @@ Required properties:
> - compatible : Should be "dlg,da9052", "dlg,da9053-aa",
> "dlg,da9053-ab", or "dlg,da9053-bb"
>
> +Optional properties:
> +- diag,tsi-as-adc : Boolean, if set the X+, X-, Y+, Y- touchscreen
> + input lines are used as general purpose analogue
> + input.
> +- diag,tsiref-microvolt: Voltage applied to the TSIREF pin, which is
> + used as reference voltage when reading TSI
> + pins as ADC. Defaults to 2500000.
> +

"dlg" is the standard vendor prefix we are using for Dialog Semiconductor
bindings. You can find this definition in the kernel file:
./Documentation/devicetree/bindings/vendor-prefixes.txt

> Sub-nodes:
> - regulators : Contain the regulator nodes. The DA9052/53 regulators are
> bound using their names as listed below:
> diff --git a/drivers/hwmon/da9052-hwmon.c b/drivers/hwmon/da9052-
> hwmon.c
> index c9832bfacfe5..a0b9d5498ba4 100644
> --- a/drivers/hwmon/da9052-hwmon.c
> +++ b/drivers/hwmon/da9052-hwmon.c
> @@ -20,13 +20,17 @@
> #include <linux/module.h>
> #include <linux/slab.h>
> #include <linux/platform_device.h>
> +#include <linux/property.h>
>
> #include <linux/mfd/da9052/da9052.h>
> #include <linux/mfd/da9052/reg.h>
>
> struct da9052_hwmon {
> - struct da9052 *da9052;
> - struct mutex hwmon_lock;
> + struct da9052 *da9052;
> + struct mutex hwmon_lock;
> + bool tsi_as_adc;
> + int tsiref_mv;
> + struct completion tsidone;
> };
>
> static const char * const input_names[] = {
> @@ -37,6 +41,10 @@ static const char * const input_names[] = {
> [DA9052_ADC_IN4] = "ADC IN4",
> [DA9052_ADC_IN5] = "ADC IN5",
> [DA9052_ADC_IN6] = "ADC IN6",
> + [DA9052_ADC_TSI_XP] = "ADC TS X+",
> + [DA9052_ADC_TSI_YP] = "ADC TS Y+",
> + [DA9052_ADC_TSI_XN] = "ADC TS X-",
> + [DA9052_ADC_TSI_YN] = "ADC TS Y-",
> [DA9052_ADC_TJUNC] = "BATTERY JUNCTION TEMP",
> [DA9052_ADC_VBBAT] = "BACK-UP BATTERY VOLTAGE",
> };
> @@ -59,6 +67,11 @@ static inline int vbbat_reg_to_mv(int value)
> return DIV_ROUND_CLOSEST(value * 5000, 1023);
> }
>
> +static inline int input_tsireg_to_mv(struct da9052_hwmon *hwmon, int
> value)
> +{
> + return DIV_ROUND_CLOSEST(value * hwmon->tsiref_mv, 1023);
> +}
> +
> static inline int da9052_enable_vddout_channel(struct da9052 *da9052)
> {
> return da9052_reg_update(da9052, DA9052_ADC_CONT_REG,
> @@ -154,6 +167,103 @@ static ssize_t da9052_read_misc_channel(struct
> device *dev,
> return sprintf(buf, "%d\n", input_reg_to_mv(ret));
> }
>
> +static int da9052_request_tsi_read(struct da9052_hwmon *hwmon, int
> channel)
> +{
> + u8 val = BIT(6); /* TSI_MAN */

There is an explicit #define for bit 6 from the TSI control register B, in the
file include/linux/mfd/da9052/reg.h #define DA9052_TSICONTB_TSIMAN 0X40

> +
> + switch (channel) {
> + case DA9052_ADC_TSI_XP:
> + val |= (0 << 4); /* TSI_MUX */
> + break;
> + case DA9052_ADC_TSI_YP:
> + val |= (1 << 4); /* TSI_MUX */
> + break;
> + case DA9052_ADC_TSI_XN:
> + val |= (2 << 4); /* TSI_MUX */
> + break;
> + case DA9052_ADC_TSI_YN:
> + val |= (3 << 4); /* TSI_MUX */
> + break;
> + }

Same for these magic numbers.
Although I don't find any entries for these in the reg.h file.

The datasheet for DA9053, Revision 2.1, says
(http://www.dialog-semiconductor.com/products/da9053)

bits (5:4) R/W TSI_MUX
Direct setting of MUX selecting which XY pin is routed to ADC_IN7 input.

00: X+ (results will be stored in TSI_XM)
01: Y+ (results will be stored in TSI_YM)
10: X- (results will be stored in TSI_XM)
11: Y- (results will be stored in TSI_YM)

May I suggest adding explicit entries to the reg.h file, somewhere near the
TSI CONTROL REGISTER B BITS line. Something like this:

#define DA9052_TSICONTB_TSIMUX_XP_XM 0x00
#define DA9052_TSICONTB_TSIMUX_YP_YM 0x10
#define DA9052_TSICONTB_TSIMUX_XN_XM 0x20
#define DA9052_TSICONTB_TSIMUX_YN_YM 0x30

> +
> + return da9052_reg_write(hwmon->da9052, DA9052_TSI_CONT_B_REG, val);
> +}
> +
> +static int da9052_get_tsi_result(struct da9052_hwmon *hwmon, int
> channel)
> +{
> + int msb, lsb;
> +
> + switch (channel) {
> + case DA9052_ADC_TSI_XP:
> + case DA9052_ADC_TSI_XN:
> + msb = da9052_reg_read(hwmon->da9052, DA9052_TSI_X_MSB_REG);
> + if (msb < 0)
> + return msb;
> +
> + lsb = da9052_reg_read(hwmon->da9052, DA9052_TSI_LSB_REG);
> + if (lsb < 0)
> + return lsb;

It seems you will risk having non-synchronous data reads by separating the MSB
and LSB read operations in this case.

Reading the datasheet on this, the TSI_X_MSB register latches the TSI_Y_MSB and
TSI_LSB registers, which in turn should allow for a single read operation across
these registers instead of four separate reads. The switch could then be
completed on the calculation instead across the four register reads.

da9052_group_read()?

> +
> + break;
> + case DA9052_ADC_TSI_YP:
> + case DA9052_ADC_TSI_YN:
> + msb = da9052_reg_read(hwmon->da9052, DA9052_TSI_Y_MSB_REG);
> + if (msb < 0)
> + return msb;
> +
> + lsb = da9052_reg_read(hwmon->da9052, DA9052_TSI_LSB_REG);
> + if (lsb < 0)
> + return lsb;
> + lsb >>= 2;

The header file does contain some symbols to access the bits in the register
fields for these MSB X/Y and LSB registers.

/* TSI X CO-ORDINATE MSB RESULT REGISTER BITS */
#define DA9052_TSIXMSB_TSIXM 0XFF

/* TSI Y CO-ORDINATE MSB RESULT REGISTER BITS */
#define DA9052_TSIYMSB_TSIYM 0XFF

/* TSI CO-ORDINATE LSB RESULT REGISTER BITS */
#define DA9052_TSILSB_PENDOWN 0X40
#define DA9052_TSILSB_TSIZL 0X30
#define DA9052_TSILSB_TSIYL 0X0C
#define DA9052_TSILSB_TSIXL 0X03

> +
> + break;
> + default:
> + return -ENXIO;

Is this a function argument error and should be, EINVAL?

> + }
> +
> + return (msb << 2) | (lsb & 0x3);
> +}

Although the reg.h register header file does not explicitly label MASK defines,
and there do not seem to be any SHIFT defines in the TSI X/Y CO-ORDINATE bits,
new entries can be added if necessary.

> +
> +
> +static ssize_t __da9052_read_tsi(struct device *dev, int channel)
> +{
> + struct da9052_hwmon *hwmon = dev_get_drvdata(dev);
> + int ret;
> +
> + reinit_completion(&hwmon->tsidone);
> +
> + ret = da9052_request_tsi_read(hwmon, channel);
> + if (ret < 0)
> + return ret;
> +
> + /* Wait for an conversion done interrupt */
> + if (!wait_for_completion_timeout(&hwmon->tsidone,
> + msecs_to_jiffies(500))) {
> + dev_err(dev, "timeout waiting for TSI conversion interrupt\n");
> + return -ETIMEDOUT;
> + }
> +
> + return da9052_get_tsi_result(hwmon, channel);
> +}
> +
> +static ssize_t da9052_read_tsi(struct device *dev,
> + struct device_attribute *devattr,
> + char *buf)
> +{
> + struct da9052_hwmon *hwmon = dev_get_drvdata(dev);
> + int channel = to_sensor_dev_attr(devattr)->index;
> + int ret;
> +
> + mutex_lock(&hwmon->hwmon_lock);
> + ret = __da9052_read_tsi(dev, channel);
> + mutex_unlock(&hwmon->hwmon_lock);
> +
> + if (ret < 0)
> + return ret;
> + else
> + return sprintf(buf, "%d\n", input_tsireg_to_mv(hwmon, ret));
> +}
> +
> static ssize_t da9052_read_tjunc(struct device *dev,
> struct device_attribute *devattr, char *buf)
> {
> @@ -196,6 +306,28 @@ static ssize_t show_label(struct device *dev,
> input_names[to_sensor_dev_attr(devattr)->index]);
> }
>
> +static umode_t da9052_channel_is_visible(struct kobject *kobj,
> + struct attribute *a, int index)
> +{
> + struct device *dev = container_of(kobj, struct device, kobj);
> + struct da9052_hwmon *hwmon = dev_get_drvdata(dev);
> +
> + switch (index) {
> + case DA9052_ADC_TSI_XP:
> + case DA9052_ADC_TSI_YP:
> + case DA9052_ADC_TSI_XN:
> + case DA9052_ADC_TSI_YN:
> + break;
> + default:
> + return a->mode;
> + }
> +
> + if (!hwmon->tsi_as_adc)
> + return 0;
> +
> + return a->mode;
> +}
> +
> static SENSOR_DEVICE_ATTR(in0_input, S_IRUGO, da9052_read_vddout,
> NULL,
> DA9052_ADC_VDDOUT);
> static SENSOR_DEVICE_ATTR(in0_label, S_IRUGO, show_label, NULL,
> @@ -221,6 +353,23 @@ static SENSOR_DEVICE_ATTR(in9_input, S_IRUGO,
> da9052_read_vbbat, NULL,
> static SENSOR_DEVICE_ATTR(in9_label, S_IRUGO, show_label, NULL,
> DA9052_ADC_VBBAT);
>
> +static SENSOR_DEVICE_ATTR(in70_input, S_IRUGO, da9052_read_tsi, NULL,
> + DA9052_ADC_TSI_XP);
> +static SENSOR_DEVICE_ATTR(in70_label, S_IRUGO, show_label, NULL,
> + DA9052_ADC_TSI_XP);
> +static SENSOR_DEVICE_ATTR(in71_input, S_IRUGO, da9052_read_tsi, NULL,
> + DA9052_ADC_TSI_XN);
> +static SENSOR_DEVICE_ATTR(in71_label, S_IRUGO, show_label, NULL,
> + DA9052_ADC_TSI_XN);
> +static SENSOR_DEVICE_ATTR(in72_input, S_IRUGO, da9052_read_tsi, NULL,
> + DA9052_ADC_TSI_YP);
> +static SENSOR_DEVICE_ATTR(in72_label, S_IRUGO, show_label, NULL,
> + DA9052_ADC_TSI_YP);
> +static SENSOR_DEVICE_ATTR(in73_input, S_IRUGO, da9052_read_tsi, NULL,
> + DA9052_ADC_TSI_YN);
> +static SENSOR_DEVICE_ATTR(in73_label, S_IRUGO, show_label, NULL,
> + DA9052_ADC_TSI_YN);
> +
> static SENSOR_DEVICE_ATTR(curr1_input, S_IRUGO, da9052_read_ich,
> NULL,
> DA9052_ADC_ICH);
> static SENSOR_DEVICE_ATTR(curr1_label, S_IRUGO, show_label, NULL,
> @@ -246,6 +395,14 @@ static struct attribute *da9052_attrs[] = {
> &sensor_dev_attr_in5_label.dev_attr.attr,
> &sensor_dev_attr_in6_input.dev_attr.attr,
> &sensor_dev_attr_in6_label.dev_attr.attr,
> + &sensor_dev_attr_in70_input.dev_attr.attr,
> + &sensor_dev_attr_in70_label.dev_attr.attr,
> + &sensor_dev_attr_in71_input.dev_attr.attr,
> + &sensor_dev_attr_in71_label.dev_attr.attr,
> + &sensor_dev_attr_in72_input.dev_attr.attr,
> + &sensor_dev_attr_in72_label.dev_attr.attr,
> + &sensor_dev_attr_in73_input.dev_attr.attr,
> + &sensor_dev_attr_in73_label.dev_attr.attr,
> &sensor_dev_attr_in9_input.dev_attr.attr,
> &sensor_dev_attr_in9_label.dev_attr.attr,
> &sensor_dev_attr_curr1_input.dev_attr.attr,
> @@ -257,29 +414,93 @@ static struct attribute *da9052_attrs[] = {
> NULL
> };
>
> -ATTRIBUTE_GROUPS(da9052);
> +static const struct attribute_group da9052_group = {
> + .attrs = da9052_attrs,
> + .is_visible = da9052_channel_is_visible,
> +};
> +__ATTRIBUTE_GROUPS(da9052);
> +
> +static irqreturn_t da9052_tsi_datardy_irq(int irq, void *data)
> +{
> + struct da9052_hwmon *hwmon = data;
> + complete(&hwmon->tsidone);
> + return IRQ_HANDLED;
> +}
>
> static int da9052_hwmon_probe(struct platform_device *pdev)
> {
> struct device *dev = &pdev->dev;
> struct da9052_hwmon *hwmon;
> struct device *hwmon_dev;
> + int err;
>
> hwmon = devm_kzalloc(dev, sizeof(struct da9052_hwmon),
> GFP_KERNEL);
> if (!hwmon)
> return -ENOMEM;
>
> + platform_set_drvdata(pdev, hwmon);
> +
> mutex_init(&hwmon->hwmon_lock);
> hwmon->da9052 = dev_get_drvdata(pdev->dev.parent);
>
> + init_completion(&hwmon->tsidone);
> +
> + hwmon->tsi_as_adc =
> + device_property_read_bool(pdev->dev.parent, "diag,tsi-as-adc");

Dialog device tree bindings should be "dlg" not "diag".

> +
> + /* get tsiref from DT, default to 2.5 Volt (typ. value in datasheet) */
> + hwmon->tsiref_mv = 2500000;
> + device_property_read_u32(pdev->dev.parent, "diag,tsiref-microvolt",
> + &hwmon->tsiref_mv);
> +
> + /* convert from microvolt (DT) to millivolt (hwmon) */
> + hwmon->tsiref_mv /= 1000;
> +
> + /* TSIREF must be between 1.8 and 2.6 Volt according to datasheet */
> + if (hwmon->tsiref_mv < 1800 || hwmon->tsiref_mv > 2600) {
> + dev_err(hwmon->da9052->dev, "invalid TSIREF voltage: %d\n",
> + hwmon->tsiref_mv);
> + return -ENXIO;
> + }
> +
> + if (hwmon->tsi_as_adc) {
> + /* disable touchscreen features */
> + da9052_reg_write(hwmon->da9052, DA9052_TSI_CONT_A_REG, 0x00);
> +
> + err = da9052_request_irq(hwmon->da9052, DA9052_IRQ_TSIREADY,
> + "tsiready-irq" da9052_tsi_datardy_irq,
> + hwmon);
> + if (err) {
> + dev_err(hwmon->da9052->dev,
> + "Failed to register TSIRDY IRQ: %d\n", err);
> + return err;
> + }
> + }
> +
> hwmon_dev = devm_hwmon_device_register_with_groups(dev, "da9052",
> hwmon,
> da9052_groups);
> - return PTR_ERR_OR_ZERO(hwmon_dev);
> + err = PTR_ERR_OR_ZERO(hwmon_dev);
> + if (err) {
> + if (hwmon->tsi_as_adc)
> + da9052_free_irq(hwmon->da9052, DA9052_IRQ_TSIREADY, hwmon);
> + return err;
> + }
> +
> + return 0;
> +}
> +
> +static int da9052_hwmon_remove(struct platform_device *pdev)
> +{
> + struct da9052_hwmon *hwmon = platform_get_drvdata(pdev);
> + if (hwmon->tsi_as_adc)
> + da9052_free_irq(hwmon->da9052, DA9052_IRQ_TSIREADY, hwmon);
> + return 0;
> }
>
> static struct platform_driver da9052_hwmon_driver = {
> .probe = da9052_hwmon_probe,
> + .remove = da9052_hwmon_remove,
> .driver = {
> .name = "da9052-hwmon",
> },
> diff --git a/drivers/input/touchscreen/da9052_tsi.c
> b/drivers/input/touchscreen/da9052_tsi.c
> index 5a013bb7bcad..f0cb239546a9 100644
> --- a/drivers/input/touchscreen/da9052_tsi.c
> +++ b/drivers/input/touchscreen/da9052_tsi.c
> @@ -16,6 +16,7 @@
> #include <linux/delay.h>
> #include <linux/platform_device.h>
> #include <linux/interrupt.h>
> +#include <linux/property.h>
>
> #include <linux/mfd/da9052/reg.h>
> #include <linux/mfd/da9052/da9052.h>
> @@ -238,6 +239,10 @@ static int da9052_ts_probe(struct platform_device
> *pdev)
> if (!da9052)
> return -EINVAL;
>
> + /* check if pins are used as general purpose ADC input */
> + if (device_property_read_bool(pdev->dev.parent, "diag,tsi-as-adc"))
> + return -ENODEV;
> +

Doesn't this presume the hwmon driver is being used?
If this "tsi-as-adc" exists in the DT binary blob and the hwmon driver is not
added, would this still disable the TS?

> tsi = kzalloc(sizeof(struct da9052_tsi), GFP_KERNEL);
> input_dev = input_allocate_device();
> if (!tsi || !input_dev) {
> diff --git a/include/linux/mfd/da9052/da9052.h
> b/include/linux/mfd/da9052/da9052.h
> index ce9230af09c2..ae5b663836d0 100644
> --- a/include/linux/mfd/da9052/da9052.h
> +++ b/include/linux/mfd/da9052/da9052.h
> @@ -45,6 +45,12 @@
> #define DA9052_ADC_TJUNC 8
> #define DA9052_ADC_VBBAT 9
>
> +/* TSI channel has its own 4 channel mux */
> +#define DA9052_ADC_TSI_XP 70
> +#define DA9052_ADC_TSI_XN 71
> +#define DA9052_ADC_TSI_YP 72
> +#define DA9052_ADC_TSI_YN 73
> +
> #define DA9052_IRQ_DCIN 0
> #define DA9052_IRQ_VBUS 1
> #define DA9052_IRQ_DCINREM 2
> --
> 2.11.0

Regards,
Steve

Next message: Keith Busch: "Re: [PATCH] fs: System memory leak when running HTX with T10 DIF enabled"
Previous message: Paul E. McKenney: "Re: [GIT PULL rcu/next] RCU commits for 4.13"
In reply to: Guenter Roeck: "Re: [PATCH 2/2] hwmon: da9052: add support for TSI channel"
Next in thread: Sebastian Reichel: "[PATCH 1/2] mfd: da9052: fix manual ADC read after timed out read"
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]