Thread (14 messages) 14 messages, 6 authors, 2026-03-27

Re: [PATCH v2 3/4] iio: adc: xilinx-xadc: Add I2C interface support

From: Jonathan Cameron <jic23@kernel.org>
Date: 2026-03-23 20:26:53
Also in: linux-devicetree, linux-iio, lkml

On Mon, 23 Mar 2026 13:15:04 +0530
Sai Krishna Potthuri [off-list ref] wrote:
Add I2C interface support for Xilinx System Management Wizard IP along
with the existing AXI memory-mapped interface. This support enables
monitoring the voltage and temperature on UltraScale+ devices where the
System Management Wizard is connected via I2C.

Key changes:
- Implement 32-bit DRP(Dynamic Reconfiguration Port) packet format as per
  Xilinx PG185 specification.
- Add separate I2C probe with xadc_i2c_of_match_table to handle same
  compatible string("xlnx,system-management-wiz-1.3") on I2C bus.
- Implement delayed version of hardware initialization for I2C interface
  to handle the case where System Management Wizard IP is not ready during
  the I2C probe.
- Add NULL checks for get_dclk_rate callback function in sampling rate
  functions to support interfaces without clock control
- Create separate iio_info structure(xadc_i2c_info) without event
  callbacks for I2C devices
- Add xadc_i2c_transaction() function to handle I2C read/write operations
- Add XADC_TYPE_US_I2C type to distinguish I2C interface from AXI

Signed-off-by: Sai Krishna Potthuri <sai.krishna.potthuri@amd.com>
Hi.
A few minor things inline.

Thanks,

Jonathan
quoted hunk ↗ jump to hunk
---
 drivers/iio/adc/Kconfig            |  15 ++
 drivers/iio/adc/Makefile           |   1 +
 drivers/iio/adc/xilinx-xadc-core.c |  28 +++-
 drivers/iio/adc/xilinx-xadc-i2c.c  | 215 +++++++++++++++++++++++++++++
 drivers/iio/adc/xilinx-xadc.h      |   1 +
 5 files changed, 256 insertions(+), 4 deletions(-)
 create mode 100644 drivers/iio/adc/xilinx-xadc-i2c.c
diff --git a/drivers/iio/adc/Kconfig b/drivers/iio/adc/Kconfig
index a4a7556f4016..5a3956a5c086 100644
--- a/drivers/iio/adc/Kconfig
+++ b/drivers/iio/adc/Kconfig
@@ -1767,6 +1767,21 @@ config XILINX_XADC
 	  The driver can also be build as a module. If so, the module will be called
 	  xilinx-xadc.
 
+config XILINX_XADC_I2C
+	tristate "Xilinx System Management Wizard I2C Interface support"
+	depends on I2C
+	select XILINX_XADC_CORE
+	help
+	  Say yes here to allow accessing the System Management
+	  Wizard on UltraScale+ devices via I2C.
+
+	  This provides voltage and temperature monitoring capabilities
+	  through the same IIO sysfs interface, but using I2C communication
+	  protocol.
+
+	  The driver can also be build as a module. If so, the module will be called
line is a little too long for Kconfig.
quoted hunk ↗ jump to hunk
+	  xilinx-xadc-i2c.
+
diff --git a/drivers/iio/adc/xilinx-xadc-i2c.c b/drivers/iio/adc/xilinx-xadc-i2c.c
new file mode 100644
index 000000000000..3d802b907260
--- /dev/null
+++ b/drivers/iio/adc/xilinx-xadc-i2c.c
@@ -0,0 +1,215 @@
+static int xadc_i2c_read_transaction(struct xadc *xadc, unsigned int reg, u16 *val)
+{
+	struct xadc_i2c *xadc_i2c = container_of(xadc, struct xadc_i2c, xadc);
+	char write_buffer[XADC_I2C_WRITE_DATA_SIZE] = { 0 };
+	struct i2c_client *client = xadc_i2c->client;
+	char read_buffer[XADC_I2C_READ_DATA_SIZE];
+	int ret;
+
+	write_buffer[2] = FIELD_GET(XADC_I2C_DRP_ADDR_MASK, reg);
+	write_buffer[3] = XADC_I2C_INSTR_READ;
+
+	ret = i2c_master_send(client, write_buffer, XADC_I2C_WRITE_DATA_SIZE);
+	if (ret < 0)
+		return ret;
+
+	ret = i2c_master_recv(client, read_buffer, XADC_I2C_READ_DATA_SIZE);
+	if (ret < 0)
+		return ret;
+
+	*val = FIELD_PREP(XADC_I2C_DRP_DATA0_MASK, read_buffer[0]) |
+	       FIELD_PREP(XADC_I2C_DRP_DATA1_MASK, read_buffer[1]);
+
+	return 0;
+}
+
+static int xadc_i2c_write_transaction(struct xadc *xadc, unsigned int reg, u16 val)
+{
+	struct xadc_i2c *xadc_i2c = container_of(xadc, struct xadc_i2c, xadc);
+	struct i2c_client *client = xadc_i2c->client;
+	char write_buffer[XADC_I2C_WRITE_DATA_SIZE];
+	int ret;
+
+	write_buffer[0] = FIELD_GET(XADC_I2C_DRP_DATA0_MASK, val);
+	write_buffer[1] = FIELD_GET(XADC_I2C_DRP_DATA1_MASK, val);
This is odd enough it might be useful to have some comments.  Why do
we need to write the value to two places for instance?
+	write_buffer[2] = FIELD_GET(XADC_I2C_DRP_ADDR_MASK, reg);
+	write_buffer[3] = XADC_I2C_INSTR_WRITE;
+
+	ret = i2c_master_send(client, write_buffer, XADC_I2C_WRITE_DATA_SIZE);
+	if (ret < 0)
+		return ret;
+
+	return 0;
+}
+
+static int xadc_hardware_init(struct xadc *xadc)
+{
+	struct xadc_i2c *xadc_i2c = container_of(xadc, struct xadc_i2c, xadc);
+	int ret;
+	u32 i;
+
+	for (i = 0; i < ARRAY_SIZE(xadc->threshold); i++) {
	for (u32 i = 0;

Though why a u32?  If size doesn't matter, convention is pretty much always
use a unsigned int for the iterator.
+		ret = xadc_i2c_read_transaction(xadc, XADC_REG_THRESHOLD(i),
+						&xadc->threshold[i]);
+		if (ret)
+			return ret;
+	}
+
+	ret = xadc_i2c_write_transaction(xadc, XADC_REG_CONF0, xadc_i2c->conf0);
+	if (ret)
+		return ret;
+
+	ret = xadc_i2c_write_transaction(xadc, XADC_REG_INPUT_MODE(0),
+					 xadc_i2c->bipolar_mask);
+	if (ret)
+		return ret;
+
+	ret = xadc_i2c_write_transaction(xadc, XADC_REG_INPUT_MODE(1),
+					 xadc_i2c->bipolar_mask >> XADC_INPUT_MODE_BITS);
+	if (ret)
+		return ret;
+
+	xadc_i2c->hw_initialized = true;
+
+	return 0;
+}
+
+static int xadc_i2c_read_reg(struct xadc *xadc, unsigned int reg, u16 *val)
+{
+	struct xadc_i2c *xadc_i2c = container_of(xadc, struct xadc_i2c, xadc);
+
+	if (!xadc_i2c->hw_initialized) {
+		int ret;
+
+		ret = xadc_hardware_init(xadc);
+		if (ret)
+			return ret;
+	}
+
+	return xadc_i2c_read_transaction(xadc, reg, val);
+}
+
+static int xadc_i2c_write_reg(struct xadc *xadc, unsigned int reg, u16 val)
+{
+	struct xadc_i2c *xadc_i2c = container_of(xadc, struct xadc_i2c, xadc);
+
+	if (!xadc_i2c->hw_initialized) {
Seems like this is always called once on first access?
If so just do it form probe and simplify the read/ write_reg() functions.
+		int ret;
+
+		ret = xadc_hardware_init(xadc);
+		if (ret)
+			return ret;
+	}
+
+	return xadc_i2c_write_transaction(xadc, reg, val);
+}
+static int xadc_i2c_probe(struct i2c_client *client)
+{
+	struct device *dev = &client->dev;
+	unsigned int conf0, bipolar_mask;
+	const struct xadc_ops *ops;
+	struct iio_dev *indio_dev;
+	struct xadc_i2c *xadc_i2c;
+	struct xadc *xadc;
+	int ret;
+
+	indio_dev = xadc_device_setup(dev, sizeof(*xadc_i2c), &ops);
+	if (IS_ERR(indio_dev))
+		return PTR_ERR(indio_dev);
+
+	xadc_i2c = iio_priv(indio_dev);
+	xadc_i2c->client = client;
+	xadc = &xadc_i2c->xadc;
+	xadc->clk = NULL;
+	xadc->ops = ops;
+	mutex_init(&xadc->mutex);
For new code (feel free to update the other code in a separate patch).
	ret = devm_mutex_init(xadc->mutex);
	if (ret)
		return ret;

As gives a small amount of lock debugging infrastructure and is now
cheap to do.  The devm form didn't used to exist.
+	spin_lock_init(&xadc->lock);
+
+	ret = xadc_device_configure(dev, indio_dev, 0, &conf0, &bipolar_mask);
+	if (ret) {
+		dev_err(dev, "Failed to setup the device: %d\n", ret);
+		return ret;
		return dev_err_probe(dev, "Failed to setup the device.\n");
Which will pretty print the error and hide it if -EPROBEDEFER (because that should
be silent) or -ENOMEM (because memory allocation errors are very noisy anyway!)
+	}
+
+	i2c_set_clientdata(client, indio_dev);
+	xadc_i2c->conf0 = conf0;
+	xadc_i2c->bipolar_mask = bipolar_mask;
+	xadc_i2c->hw_initialized = false;
+
+	return devm_iio_device_register(dev, indio_dev);
+}
  
Keyboard shortcuts
hback out one level
jnext message in thread
kprevious message in thread
ldrill in
Escclose help / fold thread tree
?toggle this help