Texas Instruments 8/10/12-bit 2/4-channel DAC driver

The DACrrcS085 (rr = 08/10/12, c = 2/4) family of SPI DACs was inherited by TI when they acquired National Semiconductor in 2011. This driver was developed for and tested with the DAC082S085 built into the Revolution Pi by KUNBUS, but should work with any of the other chips as they share the same programming interface.

There is also a family of I2C DACs with just a single channel called DACrr1C08x (rr = 08/10/12, x = 1/5). Their programming interface is very similar and it should be possible to extend the driver for these chips with moderate effort. Alternatively they could be integrated into ad5446.c. (The AD5301/AD5311/AD5321 use different power-down modes but otherwise appear to be comparable.)

Furthermore there is a family of 8-channel DACs called DACrr8S085 (rr = 08/10/12) as well as two 16-bit DACs called DAC161Sxxx (xxx = 055/997). These are more complicated devices with support for daisy-chaining and the ability to power down each channel separately. They could either be handled by a separate driver or integrated into the present driver with a larger effort.

Cc Mathias Duckeck <[email protected]>

Signed-off-by Lukas Wunner <[email protected]>

---
 Documentation/ABI/testing/sysfs-bus-iio |   1 +
 drivers/iio/dac/Kconfig                 |  10 +
 drivers/iio/dac/Makefile                |   1 +
 drivers/iio/dac/ti-dac082s085.c         | 351 ++++++++++++++++++++++++++++++++
 4 files changed, 363 insertions(+)
 create mode 100644 drivers/iio/dac/ti-dac082s085.c

Documentation/ABI/testing/sysfs-bus-iio +1 -0

index 3fc79185cc56..2e3f919485f4 100644
@@ -522,6 +522,7 @@ Description:
 		Specifies the output powerdown mode.
 		DAC output stage is disconnected from the amplifier and
 		1kohm_to_gnd: connected to ground via an 1kOhm resistor,
+		2.5kohm_to_gnd: connected to ground via a 2.5kOhm resistor,
 		6kohm_to_gnd: connected to ground via a 6kOhm resistor,
 		20kohm_to_gnd: connected to ground via a 20kOhm resistor,
 		90kohm_to_gnd: connected to ground via a 90kOhm resistor,

drivers/iio/dac/Kconfig +10 -0

index d187233dec3a..965d5c0d2468 100644
@@ -310,6 +310,16 @@ config STM32_DAC
 config STM32_DAC_CORE
 	tristate
 
+config TI_DAC082S085
+	tristate "Texas Instruments 8/10/12-bit 2/4-channel DAC driver"
+	depends on SPI_MASTER
+	help
+	  Driver for the Texas Instruments (formerly National Semiconductor)
+	  DAC082S085, DAC102S085, DAC122S085, DAC084S085, DAC104S085 and
+	  DAC124S085.
+
+	  If compiled as a module, it will be called ti-dac082s085.
+
 config VF610_DAC
 	tristate "Vybrid vf610 DAC driver"
 	depends on OF

drivers/iio/dac/Makefile +1 -0

index 8fe5af231698..4785858e04cc 100644
@@ -33,4 +33,5 @@ obj-$(CONFIG_MCP4725) += mcp4725.o
 obj-$(CONFIG_MCP4922) += mcp4922.o
 obj-$(CONFIG_STM32_DAC_CORE) += stm32-dac-core.o
 obj-$(CONFIG_STM32_DAC) += stm32-dac.o
+obj-$(CONFIG_TI_DAC082S085) += ti-dac082s085.o
 obj-$(CONFIG_VF610_DAC) += vf610_dac.o

drivers/iio/dac/ti-dac082s085.c new file +30 -0

new file mode 100644
index 000000000000..0fefb110be69
@@ -0,0 +1,351 @@
+/*
+ * ti-dac082s085.c - Texas Instruments 8/10/12-bit 2/4-channel DAC driver
+ *
+ * Copyright (C) 2017 KUNBUS GmbH
+ *
+ * http://www.ti.com/lit/ds/symlink/dac082s085.pdf
+ * http://www.ti.com/lit/ds/symlink/dac102s085.pdf
+ * http://www.ti.com/lit/ds/symlink/dac122s085.pdf
+ * http://www.ti.com/lit/ds/symlink/dac084s085.pdf
+ * http://www.ti.com/lit/ds/symlink/dac104s085.pdf
+ * http://www.ti.com/lit/ds/symlink/dac124s085.pdf
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License (version 2) as
+ * published by the Free Software Foundation.
+ */
+
+#include <linux/iio/iio.h>
+#include <linux/module.h>
+#include <linux/regulator/consumer.h>
+#include <linux/spi/spi.h>
+
+/**
+ * struct ti_dac_chip - TI DAC chip
+ * @lock: protects write sequences
+ * @vref: regulator generating Vref
+ * @mesg: SPI message to perform a write
+ * @xfer: SPI transfer used by @mesg
+ * @buf: buffer for @xfer
+ * @val: cached value of each output

drivers/iio/dac/ti-dac082s085.c new file +311 -0

+	u16 val[4];
+	bool powerdown;
+	u8 powerdown_mode;
+	u8 resolution;
+};
+
+#define WRITE_NOT_UPDATE(chan)	(0x00 | (chan) << 6)
+#define WRITE_AND_UPDATE(chan)	(0x10 | (chan) << 6)
+#define WRITE_ALL_UPDATE	 0x20
+#define POWERDOWN(mode) 	(0x30 | ((mode) + 1) << 6)
+
+static int ti_dac_cmd(struct ti_dac_chip *ti_dac, u8 cmd, u16 val)
+{
+	u8 shift = 12 - ti_dac->resolution;
+
+	ti_dac->buf[0] = cmd | (val >> (8 - shift));
+	ti_dac->buf[1] = (val << shift) & 0xff;
+	return spi_sync(ti_dac->mesg.spi, &ti_dac->mesg);
+}
+
+static const char * const ti_dac_powerdown_modes[] = {
+	"2.5kohm_to_gnd", "100kohm_to_gnd", "three_state",
+};
+
+static int ti_dac_get_powerdown_mode(struct iio_dev *indio_dev,
+				     const struct iio_chan_spec *chan)
+{
+	struct ti_dac_chip *ti_dac = iio_priv(indio_dev);
+
+	return ti_dac->powerdown_mode;
+}
+
+static int ti_dac_set_powerdown_mode(struct iio_dev *indio_dev,
+				     const struct iio_chan_spec *chan,
+				     unsigned int mode)
+{
+	struct ti_dac_chip *ti_dac = iio_priv(indio_dev);
+	int ret = 0;
+
+	if (ti_dac->powerdown_mode == mode)
+		return 0;
+
+	mutex_lock(&ti_dac->lock);
+	if (ti_dac->powerdown) {
+		ret = ti_dac_cmd(ti_dac, POWERDOWN(mode), 0);
+		if (ret)
+			goto out;
+	}
+	ti_dac->powerdown_mode = mode;
+
+out:
+	mutex_unlock(&ti_dac->lock);
+	return ret;
+}
+
+static const struct iio_enum ti_dac_powerdown_mode = {
+	.items = ti_dac_powerdown_modes,
+	.num_items = ARRAY_SIZE(ti_dac_powerdown_modes),
+	.get = ti_dac_get_powerdown_mode,
+	.set = ti_dac_set_powerdown_mode,
+};
+
+static ssize_t ti_dac_read_powerdown(struct iio_dev *indio_dev,
+				     uintptr_t private,
+				     const struct iio_chan_spec *chan,
+				     char *buf)
+{
+	struct ti_dac_chip *ti_dac = iio_priv(indio_dev);
+
+	return sprintf(buf, "%d\n", ti_dac->powerdown);
+}
+
+static ssize_t ti_dac_write_powerdown(struct iio_dev *indio_dev,
+				      uintptr_t private,
+				      const struct iio_chan_spec *chan,
+				      const char *buf, size_t len)
+{
+	struct ti_dac_chip *ti_dac = iio_priv(indio_dev);
+	bool powerdown;
+	int ret;
+
+	ret = strtobool(buf, &powerdown);
+	if (ret)
+		return ret;
+
+	if (ti_dac->powerdown == powerdown)
+		return len;
+
+	mutex_lock(&ti_dac->lock);
+	if (powerdown)
+		ret = ti_dac_cmd(ti_dac, POWERDOWN(ti_dac->powerdown_mode), 0);
+	else
+		ret = ti_dac_cmd(ti_dac, WRITE_AND_UPDATE(0), ti_dac->val[0]);
+	if (!ret)
+		ti_dac->powerdown = powerdown;
+	mutex_unlock(&ti_dac->lock);
+
+	return ret ? ret : len;
+}
+
+static const struct iio_chan_spec_ext_info ti_dac_ext_info[] = {
+	{
+		.name	   = "powerdown",
+		.read	   = ti_dac_read_powerdown,
+		.write	   = ti_dac_write_powerdown,
+		.shared	   = IIO_SHARED_BY_TYPE,
+	},
+	IIO_ENUM("powerdown_mode", IIO_SHARED_BY_TYPE, &ti_dac_powerdown_mode),
+	IIO_ENUM_AVAILABLE("powerdown_mode", &ti_dac_powerdown_mode),
+	{ },
+};
+
+#define TI_DAC_CHANNEL(chan) {					\
+	.type = IIO_VOLTAGE,					\
+	.channel = (chan),					\
+	.address = (chan),					\
+	.indexed = true,					\
+	.output = true,						\
+	.datasheet_name = (const char[]){ 'A' + (chan), 0 },	\
+	.info_mask_separate = BIT(IIO_CHAN_INFO_RAW),		\
+	.info_mask_shared_by_type = BIT(IIO_CHAN_INFO_SCALE),	\
+	.ext_info = ti_dac_ext_info,				\
+}
+
+static const struct iio_chan_spec ti_dac_channels[] = {
+	TI_DAC_CHANNEL(0),
+	TI_DAC_CHANNEL(1),
+	TI_DAC_CHANNEL(2),
+	TI_DAC_CHANNEL(3),
+};
+
+static int ti_dac_read_raw(struct iio_dev *indio_dev,
+			   struct iio_chan_spec const *chan,
+			   int *val, int *val2, long mask)
+{
+	struct ti_dac_chip *ti_dac = iio_priv(indio_dev);
+	int ret;
+
+	switch (mask) {
+	case IIO_CHAN_INFO_RAW:
+		*val = ti_dac->val[chan->channel];
+		ret = IIO_VAL_INT;
+		break;
+
+	case IIO_CHAN_INFO_SCALE:
+		ret = regulator_get_voltage(ti_dac->vref);
+		if (ret < 0)
+			return ret;
+
+		*val = ret / 1000;
+		*val2 = ti_dac->resolution;
+		ret = IIO_VAL_FRACTIONAL_LOG2;
+		break;
+
+	default:
+		ret = -EINVAL;
+	}
+
+	return ret;
+}
+
+static int ti_dac_write_raw(struct iio_dev *indio_dev,
+			    struct iio_chan_spec const *chan,
+			    int val, int val2, long mask)
+{
+	struct ti_dac_chip *ti_dac = iio_priv(indio_dev);
+	int ret;
+
+	switch (mask) {
+	case IIO_CHAN_INFO_RAW:
+		if (ti_dac->val[chan->channel] == val)
+			return 0;
+
+		if (val >= (1 << ti_dac->resolution) || val < 0)
+			return -EINVAL;
+
+		if (ti_dac->powerdown)
+			return -EBUSY;
+
+		mutex_lock(&ti_dac->lock);
+		ret = ti_dac_cmd(ti_dac, WRITE_AND_UPDATE(chan->channel), val);
+		if (!ret)
+			ti_dac->val[chan->channel] = val;
+		mutex_unlock(&ti_dac->lock);
+		break;
+
+	default:
+		ret = -EINVAL;
+	}
+
+	return ret;
+}
+
+static int ti_dac_write_raw_get_fmt(struct iio_dev *indio_dev,
+				    struct iio_chan_spec const *chan, long mask)
+{
+	return IIO_VAL_INT;
+}
+
+static const struct iio_info ti_dac_info = {
+	.read_raw	   = ti_dac_read_raw,
+	.write_raw	   = ti_dac_write_raw,
+	.write_raw_get_fmt = ti_dac_write_raw_get_fmt,
+};
+
+static int ti_dac_probe(struct spi_device *spi)
+{
+	struct device *dev = &spi->dev;
+	struct ti_dac_chip *ti_dac;
+	struct iio_dev *indio_dev;
+	int ret;
+
+	indio_dev = devm_iio_device_alloc(dev, sizeof(*ti_dac));
+	if (!indio_dev)
+		return -ENOMEM;
+
+	indio_dev->dev.parent = dev;
+	indio_dev->info = &ti_dac_info;
+	indio_dev->name = spi->modalias;
+	indio_dev->modes = INDIO_DIRECT_MODE;
+	indio_dev->channels = ti_dac_channels;
+	spi_set_drvdata(spi, indio_dev);
+
+	ti_dac = iio_priv(indio_dev);
+	ti_dac->xfer.tx_buf = &ti_dac->buf;
+	ti_dac->xfer.len = sizeof(ti_dac->buf);
+	spi_message_init_with_transfers(&ti_dac->mesg, &ti_dac->xfer, 1);
+	ti_dac->mesg.spi = spi;
+
+	ret = sscanf(spi->modalias, "dac%2hhu%1d",
+		     &ti_dac->resolution, &indio_dev->num_channels);
+	WARN_ON(ret != 2);
+
+	ti_dac->vref = devm_regulator_get(dev, "vref");
+	if (IS_ERR(ti_dac->vref))
+		return PTR_ERR(ti_dac->vref);
+
+	ret = regulator_enable(ti_dac->vref);
+	if (ret < 0)
+		return ret;
+
+	mutex_init(&ti_dac->lock);
+
+	ret = ti_dac_cmd(ti_dac, WRITE_ALL_UPDATE, 0);
+	if (ret) {
+		dev_err(dev, "failed to initialize outputs to 0\n");
+		goto err;
+	}
+
+	ret = iio_device_register(indio_dev);
+	if (ret)
+		goto err;
+
+	return 0;
+
+err:
+	mutex_destroy(&ti_dac->lock);
+	regulator_disable(ti_dac->vref);
+	return ret;
+}
+
+static int ti_dac_remove(struct spi_device *spi)
+{
+	struct iio_dev *indio_dev = spi_get_drvdata(spi);
+	struct ti_dac_chip *ti_dac = iio_priv(indio_dev);
+
+	iio_device_unregister(indio_dev);
+	mutex_destroy(&ti_dac->lock);
+	regulator_disable(ti_dac->vref);
+
+	return 0;
+}
+
+#ifdef CONFIG_OF
+static const struct of_device_id ti_dac_of_id[] = {
+	{ .compatible = "ti,dac082s085" },
+	{ .compatible = "ti,dac102s085" },
+	{ .compatible = "ti,dac122s085" },
+	{ .compatible = "ti,dac084s085" },
+	{ .compatible = "ti,dac104s085" },
+	{ .compatible = "ti,dac124s085" },
+	{ }
+};
+MODULE_DEVICE_TABLE(of, ti_dac_of_id);
+#endif
+
+static const struct spi_device_id ti_dac_spi_id[] = {
+	{ "dac082s085" },
+	{ "dac102s085" },
+	{ "dac122s085" },
+	{ "dac084s085" },
+	{ "dac104s085" },
+	{ "dac124s085" },
+	{ }
+};
+MODULE_DEVICE_TABLE(spi, ti_dac_spi_id);
+
+static struct spi_driver ti_dac_driver = {
+	.driver = {
+		.name		= "ti-dac082s085",
+		.of_match_table	= of_match_ptr(ti_dac_of_id),
+	},
+	.probe	  = ti_dac_probe,
+	.remove   = ti_dac_remove,
+	.id_table = ti_dac_spi_id,
+};
+module_spi_driver(ti_dac_driver);
+
+MODULE_AUTHOR("Lukas Wunner <[email protected]>");
+MODULE_DESCRIPTION("Texas Instruments 8/10/12-bit 2/4-channel DAC driver");
+MODULE_LICENSE("GPL v2");

The two properties unique to each supported chip, resolution and number of channels, are currently gleaned from the chip's name. E.g. dac102s085 is a dual channel 10-bit DAC.

        ^^^

This was deemed unmaintainable by the subsystem maintainer once the driver is extended to support further chips, hence it was requested to add an explicit table for chip-specific information and use an enum to reference into it.

This adds 17 LoC without any immediate gain, so make the change in a separate commit which can be reverted if we determine in 10 years that it was unnecessary.

Signed-off-by Lukas Wunner <[email protected]>

---
 drivers/iio/dac/ti-dac082s085.c | 35 ++++++++++++++++++++++++++---------
 1 file changed, 26 insertions(+), 9 deletions(-)

drivers/iio/dac/ti-dac082s085.c +26 -9

index 0fefb110be69..2c1556f32574 100644
@@ -20,6 +20,22 @@
 #include <linux/regulator/consumer.h>
 #include <linux/spi/spi.h>
 
+enum { dual_8bit, dual_10bit, dual_12bit, quad_8bit, quad_10bit, quad_12bit };
+
+struct ti_dac_spec {
+	u8 num_channels;
+	u8 resolution;
+};
+
+static const struct ti_dac_spec ti_dac_spec[] = {
+	[dual_8bit]  = { .num_channels = 2, .resolution = 8  },
+	[dual_10bit] = { .num_channels = 2, .resolution = 10 },
+	[dual_12bit] = { .num_channels = 2, .resolution = 12 },
+	[quad_8bit]  = { .num_channels = 4, .resolution = 8  },
+	[quad_10bit] = { .num_channels = 4, .resolution = 10 },
+	[quad_12bit] = { .num_channels = 4, .resolution = 12 },
+};
+
 /**
  * struct ti_dac_chip - TI DAC chip
  * @lock: protects write sequences

@@ -246,6 +262,7 @@ static const struct iio_info ti_dac_info = {
 static int ti_dac_probe(struct spi_device *spi)
 {
 	struct device *dev = &spi->dev;
+	const struct ti_dac_spec *spec;
 	struct ti_dac_chip *ti_dac;
 	struct iio_dev *indio_dev;
 	int ret;

@@ -267,9 +284,9 @@ static int ti_dac_probe(struct spi_device *spi)
 	spi_message_init_with_transfers(&ti_dac->mesg, &ti_dac->xfer, 1);
 	ti_dac->mesg.spi = spi;
 
-	ret = sscanf(spi->modalias, "dac%2hhu%1d",
-		     &ti_dac->resolution, &indio_dev->num_channels);
-	WARN_ON(ret != 2);
+	spec = &ti_dac_spec[spi_get_device_id(spi)->driver_data];
+	indio_dev->num_channels = spec->num_channels;
+	ti_dac->resolution = spec->resolution;
 
 	ti_dac->vref = devm_regulator_get(dev, "vref");
 	if (IS_ERR(ti_dac->vref))

@@ -325,12 +342,12 @@ MODULE_DEVICE_TABLE(of, ti_dac_of_id);
 #endif
 
 static const struct spi_device_id ti_dac_spi_id[] = {
-	{ "dac082s085" },
-	{ "dac102s085" },
-	{ "dac122s085" },
-	{ "dac084s085" },
-	{ "dac104s085" },
-	{ "dac124s085" },
+	{ "dac082s085", dual_8bit  },
+	{ "dac102s085", dual_10bit },
+	{ "dac122s085", dual_12bit },
+	{ "dac084s085", quad_8bit  },
+	{ "dac104s085", quad_10bit },
+	{ "dac124s085", quad_12bit },
 	{ }
 };
 MODULE_DEVICE_TABLE(spi, ti_dac_spi_id);

Add device tree bindings for Texas Instruments 8/10/12-bit 2/4-channel DAC driver.

Cc Mathias Duckeck <[email protected]>

Acked-by Rob Herring <[email protected]>

Signed-off-by Lukas Wunner <[email protected]>

---
 .../devicetree/bindings/iio/dac/ti-dac082s085.txt  | 34 ++++++++++++++++++++++
 1 file changed, 34 insertions(+)
 create mode 100644 Documentation/devicetree/bindings/iio/dac/ti-dac082s085.txt

Documentation/devicetree/bindings/iio/dac/ti-dac082s085.txt new file +34 -0

new file mode 100644
index 000000000000..9cb0e10df704
@@ -0,0 +1,34 @@
+Texas Instruments 8/10/12-bit 2/4-channel DAC driver
+
+Required properties:
+ - compatible:		Must be one of:
+			"ti,dac082s085"
+			"ti,dac102s085"
+			"ti,dac122s085"
+			"ti,dac084s085"
+			"ti,dac104s085"
+			"ti,dac124s085"
+ - reg: 		Chip select number.
+ - spi-cpha, spi-cpol:	SPI mode (0,1) or (1,0) must be used, so specify
+			either spi-cpha or spi-cpol (but not both).
+ - vref-supply: 	Phandle to the external reference voltage supply.
+
+For other required and optional properties of SPI slave nodes please refer to
+../../spi/spi-bus.txt.
+
+Example:
+	vref_2v5_reg: regulator-vref {
+		compatible = "regulator-fixed";
+		regulator-name = "2v5";
+		regulator-min-microvolt = <2500000>;
+		regulator-max-microvolt = <2500000>;
+		regulator-always-on;
+	};
+
+	dac@0 {
+		compatible = "ti,dac082s085";
+		reg = <0>;
+		spi-max-frequency = <40000000>;
+		spi-cpol;
+		vref-supply = <&vref_2v5_reg>;
+	};

drivers/iio/dac/ti-dac082s085.c line 40

Quoted 6 lines · On Tue, 17 Oct 2017 12:42:00 +0200 Lukas Wunner <[email protected]> wrote:

drivers/iio/dac/ti-dac082s085.c new file +6 -0

+struct ti_dac_chip {
+	struct mutex lock;
+	struct regulator *vref;
+	struct spi_message mesg;
+	struct spi_transfer xfer;
+	u8 buf[2] ____cacheline_aligned;

I missed this the first time. The whole point of the fun of ____cacheline_aligned is to ensure nothing shares a cacheline with the buffer used for DMA. We take care when allocating these private structures that this will be true, but it relies on nothing that might be changed during dma being after the buffer.

Simple reordering fix so I'll do it whilst applying.

Quoted 15 lines · Jonathan Cameron

drivers/iio/dac/ti-dac082s085.c line 40

Quoted 6 lines · On Tue, 17 Oct 2017 12:42:00 +0200 Lukas Wunner <[email protected]> wrote:

drivers/iio/dac/ti-dac082s085.c new file +6 -0

+struct ti_dac_chip {
+	struct mutex lock;
+	struct regulator *vref;
+	struct spi_message mesg;
+	struct spi_transfer xfer;
+	u8 buf[2] ____cacheline_aligned;  

I missed this the first time. The whole point of the fun of ____cacheline_aligned is to ensure nothing shares a cacheline with the buffer used for DMA. We take care when allocating these private structures that this will be true, but it relies on nothing that might be changed during dma being after the buffer.

Quoted 1 line

Simple reordering fix so I'll do it whilst applying.

TBH I was a bit confused about how to achieve proper cacheline alignment. My limited understanding of the slab allocator is that it neither guarantees that an allocation begins on a cacheline nor that no another allocation immediately starts after it.

However taking a closer look at iio_device_alloc() I realize now that it implicitly takes care of both by inserting the necessary padding. Neat. I was already wondering why the IIO subsystem takes this somewhat unusual approach to allocate the private struct behind struct iio_dev, instead of letting the driver embed it within its private struct and use container_of() to yield the private struct corresponding to a struct iio_dev, as most other subsystems do. But that explains it.

Thanks,

Lukas