Re: [PATCH v3 1/2] leds: ltc3208: add driver
From: Lee Jones
Date: Thu Apr 09 2026 - 12:59:13 EST
"Add driver" is not a good subject line.
> Kernel driver implementation for LTC3208 Multidisplay LED Driver
A one line commit messages is not suitable fore a 300 line driver!
What is the LTC3208 Multidisplay LED?
What does it do?
How does it operate?
What's special about it?
Any quirks?
> Signed-off-by: Jan Carlo Roleda <jancarlo.roleda@xxxxxxxxxx>
> ---
> MAINTAINERS | 7 ++
> drivers/leds/Kconfig | 11 ++
> drivers/leds/Makefile | 1 +
> drivers/leds/leds-ltc3208.c | 298 ++++++++++++++++++++++++++++++++++++++++++++
> 4 files changed, 317 insertions(+)
>
> diff --git a/MAINTAINERS b/MAINTAINERS
> index 55af015174a5..48bae02057d5 100644
> --- a/MAINTAINERS
> +++ b/MAINTAINERS
> @@ -15126,6 +15126,13 @@ W: https://ez.analog.com/linux-software-drivers
> F: Documentation/devicetree/bindings/iio/temperature/adi,ltc2983.yaml
> F: drivers/iio/temperature/ltc2983.c
>
> +LTC3208 LED DRIVER
> +M: Jan Carlo Roleda <jancarlo.roleda@xxxxxxxxxx>
> +L: linux-leds@xxxxxxxxxxxxxxx
> +S: Maintained
> +W: https://ez.analog.com/linux-software-drivers
> +F: drivers/leds/leds-ltc3208.c
> +
> LTC4282 HARDWARE MONITOR DRIVER
> M: Nuno Sa <nuno.sa@xxxxxxxxxx>
> L: linux-hwmon@xxxxxxxxxxxxxxx
> diff --git a/drivers/leds/Kconfig b/drivers/leds/Kconfig
> index 597d7a79c988..867b120ea8ba 100644
> --- a/drivers/leds/Kconfig
> +++ b/drivers/leds/Kconfig
> @@ -1029,6 +1029,17 @@ config LEDS_ACER_A500
> This option enables support for the Power Button LED of
> Acer Iconia Tab A500.
>
> +config LEDS_LTC3208
> + tristate "LED Driver for Analog Devices LTC3208"
> + depends on LEDS_CLASS && I2C
> + select REGMAP_I2C
> + help
> + Say Y to enable the LTC3208 LED driver.
> + This supports the LED device LTC3208.
You can do better!
> + To compile this driver as a module, choose M here: the module will
> + be called ltc3208.
> +
> source "drivers/leds/blink/Kconfig"
>
> comment "Flash and Torch LED drivers"
> diff --git a/drivers/leds/Makefile b/drivers/leds/Makefile
> index 8fdb45d5b439..b08b539112b6 100644
> --- a/drivers/leds/Makefile
> +++ b/drivers/leds/Makefile
> @@ -61,6 +61,7 @@ obj-$(CONFIG_LEDS_LP8788) += leds-lp8788.o
> obj-$(CONFIG_LEDS_LP8860) += leds-lp8860.o
> obj-$(CONFIG_LEDS_LP8864) += leds-lp8864.o
> obj-$(CONFIG_LEDS_LT3593) += leds-lt3593.o
> +obj-$(CONFIG_LEDS_LTC3208) += leds-ltc3208.o
> obj-$(CONFIG_LEDS_MAX5970) += leds-max5970.o
> obj-$(CONFIG_LEDS_MAX77650) += leds-max77650.o
> obj-$(CONFIG_LEDS_MAX77705) += leds-max77705.o
> diff --git a/drivers/leds/leds-ltc3208.c b/drivers/leds/leds-ltc3208.c
> new file mode 100644
> index 000000000000..65e65cd73d73
> --- /dev/null
> +++ b/drivers/leds/leds-ltc3208.c
> @@ -0,0 +1,298 @@
> +// SPDX-License-Identifier: GPL-2.0
> +/*
> + * LED driver for Analog Devices LTC3208 Multi-Display Driver
> + *
> + * Copyright 2026 Analog Devices Inc.
> + *
> + * Author: Jan Carlo Roleda <jancarlo.roleda@xxxxxxxxxx>
> + */
> +#include <linux/bitfield.h>
> +#include <linux/errno.h>
> +#include <linux/gpio/consumer.h>
> +#include <linux/i2c.h>
> +#include <linux/init.h>
> +#include <linux/leds.h>
> +#include <linux/mod_devicetable.h>
> +#include <linux/module.h>
> +#include <linux/platform_device.h>
> +#include <linux/property.h>
> +#include <linux/regmap.h>
> +#include <linux/types.h>
> +#include <linux/workqueue.h>
Are all of these headers strictly necessary? For instance, it doesn't appear we
are using GPIOs, platform devices, or workqueues in this driver.
> +#define LTC3208_SET_HIGH_BYTE_DATA(x) FIELD_PREP(GENMASK(7, 4), (x))
> +
> +/* Registers */
> +#define LTC3208_REG_A_GRNRED 0x1 /* Green (High half-byte) and Red (Low half-byte) current DAC*/
> +#define LTC3208_REG_B_AUXBLU 0x2 /* AUX (High half-byte) and Blue (Low half-byte) current DAC*/
> +#define LTC3208_REG_C_MAIN 0x3 /* Main current DAC */
> +#define LTC3208_REG_D_SUB 0x4 /* Sub current DAC */
> +#define LTC3208_REG_E_AUX 0x5 /* AUX DAC Select */
> +#define LTC3208_REG_F_CAM 0x6 /* CAM (High half-byte and Low half-byte) current DAC*/
> +#define LTC3208_REG_G_OPT 0x7 /* Device Options */
> +
> +/* Device Options register */
> +#define LTC3208_OPT_CPO_MASK GENMASK(7, 6)
> +#define LTC3208_OPT_DIS_RGBDROP BIT(3)
> +#define LTC3208_OPT_DIS_CAMHILO BIT(2)
> +#define LTC3208_OPT_EN_RGBS BIT(1)
Nit: This can look nicer nested:
#define LTC3208_REG_A_GRNRED 0x1 /* Green (High half-byte) and Red (Low half-byte) current DAC*/
#define LTC3208_REG_B_AUXBLU 0x2 /* AUX (High half-byte) and Blue (Low half-byte) current DAC*/
#define LTC3208_REG_C_MAIN 0x3 /* Main current DAC */
#define LTC3208_REG_D_SUB 0x4 /* Sub current DAC */
#define LTC3208_REG_E_AUX 0x5 /* AUX DAC Select */
#define LTC3208_AUX1_MASK GENMASK(1, 0)
#define LTC3208_AUX2_MASK GENMASK(3, 2)
#define LTC3208_AUX3_MASK GENMASK(5, 4)
#define LTC3208_AUX4_MASK GENMASK(7, 6)
#define LTC3208_REG_F_CAM 0x6 /* CAM (High half-byte and Low half-byte) current DAC*/
#define LTC3208_REG_G_OPT 0x7 /* Device Options */
#define LTC3208_OPT_CPO_MASK GENMASK(7, 6)
#define LTC3208_OPT_DIS_RGBDROP BIT(3)
#define LTC3208_OPT_DIS_CAMHILO BIT(2)
#define LTC3208_OPT_EN_RGBS BIT(1)
> +#define LTC3208_MAX_BRIGHTNESS_4BIT 0xF
> +#define LTC3208_MAX_BRIGHTNESS_8BIT 0xFF
> +
> +#define LTC3208_NUM_LED_GRPS 8
> +#define LTC3208_NUM_AUX_LEDS 4
> +
> +#define LTC3208_NUM_AUX_OPT 4
> +#define LTC3208_MAX_CPO_OPT 3
Nit: Can we have _all_ of the values line up nicely?
#define LTC3208_NUM_AUX_OPT 4
#define LTC3208_MAX_CPO_OPT 3
> +enum ltc3208_aux_channel {
> + LTC3208_AUX_CHAN_AUX = 0,
> + LTC3208_AUX_CHAN_MAIN,
> + LTC3208_AUX_CHAN_SUB,
> + LTC3208_AUX_CHAN_CAM
> +};
> +
> +enum ltc3208_channel {
> + LTC3208_CHAN_MAIN = 0,
> + LTC3208_CHAN_SUB,
> + LTC3208_CHAN_AUX,
> + LTC3208_CHAN_CAML,
> + LTC3208_CHAN_CAMH,
> + LTC3208_CHAN_RED,
> + LTC3208_CHAN_BLUE,
> + LTC3208_CHAN_GREEN
> +};
> +
> +static const char * const ltc3208_dt_aux_channels[] = {
> + "adi,aux1-channel", "adi,aux2-channel",
> + "adi,aux3-channel", "adi,aux4-channel"
> +};
> +
> +static const char * const ltc3208_aux_opt[] = {
> + "aux", "main", "sub", "cam"
> +};
> +
> +
?
> +struct ltc3208_led {
> + struct led_classdev cdev;
> + struct i2c_client *client;
> + enum ltc3208_channel channel;
> +};
> +
> +struct ltc3208_dev {
> + struct i2c_client *client;
> + struct regmap *map;
> + struct ltc3208_led *leds;
> +};
> +
> +static const struct regmap_config ltc3208_regmap_cfg = {
> + .reg_bits = 8,
> + .val_bits = 8,
> +};
> +
> +static int ltc3208_led_set_brightness(struct led_classdev *led_cdev,
> + enum led_brightness brightness)
> +{
> + struct ltc3208_led *led = container_of(led_cdev,
> + struct ltc3208_led, cdev);
You can use 100-chars to avoid this awkwardness.
> + struct i2c_client *client = led->client;
> + struct ltc3208_dev *dev = i2c_get_clientdata(client);
> + struct regmap *map = dev->map;
> + u8 current_level = brightness;
> +
> + /*
> + * For registers with 4-bit splits (CAM, AUX/BLUE, GREEN/RED), the other
> + * half of the byte will be retrieved from the stored DAC value before
> + * updating the register.
> + */
> + switch (led->channel) {
> + case LTC3208_CHAN_MAIN:
> + return regmap_write(map, LTC3208_REG_C_MAIN, current_level);
> + case LTC3208_CHAN_SUB:
> + return regmap_write(map, LTC3208_REG_D_SUB, current_level);
> + case LTC3208_CHAN_AUX:
> + /* combine both low and high halves of byte */
> + current_level = LTC3208_SET_HIGH_BYTE_DATA(current_level);
> + current_level |= dev->leds[LTC3208_CHAN_BLUE].cdev.brightness;
> + return regmap_write(map, LTC3208_REG_B_AUXBLU, current_level);
Should we be using 'regmap_update_bits()' or 'regmap_field' here instead?
Constructing the register value by reading the software state of another LED
instance could lead to races.
> + case LTC3208_CHAN_BLUE:
> + /* apply high bits stored in other led */
> + current_level |= LTC3208_SET_HIGH_BYTE_DATA(dev->leds[LTC3208_CHAN_AUX].cdev.brightness);
> + return regmap_write(map, LTC3208_REG_B_AUXBLU, current_level);
> + case LTC3208_CHAN_CAMH:
> + current_level = LTC3208_SET_HIGH_BYTE_DATA(current_level);
> + current_level |= dev->leds[LTC3208_CHAN_CAML].cdev.brightness;
> + return regmap_write(map, LTC3208_REG_F_CAM, current_level);
> + case LTC3208_CHAN_CAML:
> + current_level |= LTC3208_SET_HIGH_BYTE_DATA(dev->leds[LTC3208_CHAN_CAMH].cdev.brightness);
> + return regmap_write(map, LTC3208_REG_F_CAM, current_level);
> + case LTC3208_CHAN_GREEN:
> + current_level = LTC3208_SET_HIGH_BYTE_DATA(current_level);
> + current_level |= dev->leds[LTC3208_CHAN_RED].cdev.brightness;
> + return regmap_write(map, LTC3208_REG_A_GRNRED, current_level);
> + case LTC3208_CHAN_RED:
> + current_level |= LTC3208_SET_HIGH_BYTE_DATA(dev->leds[LTC3208_CHAN_GREEN].cdev.brightness);
> + return regmap_write(map, LTC3208_REG_A_GRNRED, current_level);
This lot is begging for a sub function:
static int ltc3208_led_set_current_level(struct regmap *regmap, u8 reg, u8 low, u8 high) {
{
return regmap_write(regmap, reg, SET_HIGH_BYTE(high) | low);
}
> + default:
> + dev_err(&client->dev, "Invalid LED Channel\n");
> + return -EINVAL;
> + }
> +}
> +
> +static int ltc3208_update_options(struct ltc3208_dev *dev,
> + bool is_sub, bool is_cam_hi, bool is_rgb_drop)
> +{
> + struct regmap *map = dev->map;
> + u8 val = FIELD_PREP(LTC3208_OPT_EN_RGBS, is_sub) |
> + FIELD_PREP(LTC3208_OPT_DIS_CAMHILO, is_cam_hi) |
> + FIELD_PREP(LTC3208_OPT_DIS_RGBDROP, is_rgb_drop);
> +
That tabbing is awkward. In these cases it's better to do the
allocation after the declaration.
> + return regmap_write(map, LTC3208_REG_G_OPT, val);
> +}
> +
> +static int ltc3208_update_aux_dac(struct ltc3208_dev *dev,
> + enum ltc3208_aux_channel aux_1, enum ltc3208_aux_channel aux_2,
> + enum ltc3208_aux_channel aux_3, enum ltc3208_aux_channel aux_4)
These should sit under the '('.
> +{
> + struct regmap *map = dev->map;
> + u8 val = FIELD_PREP(LTC3208_AUX1_MASK, aux_1) |
> + FIELD_PREP(LTC3208_AUX2_MASK, aux_2) |
> + FIELD_PREP(LTC3208_AUX3_MASK, aux_3) |
> + FIELD_PREP(LTC3208_AUX4_MASK, aux_4);
As above.
> + return regmap_write(map, LTC3208_REG_E_AUX, val);
> +}
> +
> +static int ltc3208_probe(struct i2c_client *client)
> +{
> + enum ltc3208_aux_channel aux_channels[LTC3208_NUM_AUX_LEDS];
> + struct ltc3208_dev *data;
'data' is a terrible variable name.
> + struct ltc3208_led *leds;
> + struct regmap *map;
'regmap'
> + int ret, i;
> + u32 val;
> + bool dropdis_rgb_aux4;
> + bool dis_camhl;
> + bool en_rgbs;
> +
> + map = devm_regmap_init_i2c(client, <c3208_regmap_cfg);
> + if (IS_ERR(map))
> + return dev_err_probe(&client->dev, PTR_ERR(map),
> + "Failed to initialize regmap\n");
> +
> + data = devm_kzalloc(&client->dev, sizeof(*data), GFP_KERNEL);
> + if (!data)
> + return -ENOMEM;
> +
> + leds = devm_kcalloc(&client->dev, LTC3208_NUM_LED_GRPS,
> + sizeof(struct ltc3208_led), GFP_KERNEL);
> + if (!leds)
> + return -ENOMEM;
> +
> + data->client = client;
> + data->map = map;
> +
> + /* initialize options from devicetree */
Capitalise comments and it's "Device Tree", although honestly, I think
the whole comment is superfluous.
> + dis_camhl = device_property_read_bool(&client->dev,
> + "adi,disable-camhl-pin");
> + en_rgbs = device_property_read_bool(&client->dev,
> + "adi,cfg-enrgbs-pin");
> + dropdis_rgb_aux4 = device_property_read_bool(&client->dev,
> + "adi,disable-rgb-aux4-dropout");
Use 100-chars.
> + ret = ltc3208_update_options(data, en_rgbs, dis_camhl,
> + dropdis_rgb_aux4);
> + if (ret)
> + return dev_err_probe(&client->dev, ret,
> + "error writing to options register\n");
Capitalise.
> + /* initialize aux channel configurations from devicetree */
As above and throughout.
> + for (i = 0; i < LTC3208_NUM_AUX_LEDS; i++) {
for (int i = 0; ...
> + ret = device_property_match_property_string(&client->dev,
> + ltc3208_dt_aux_channels[i],
> + ltc3208_aux_opt,
> + LTC3208_NUM_AUX_OPT);
> + /* use default value if absent in devicetree */
> + if (ret == -EINVAL)
> + aux_channels[i] = LTC3208_AUX_CHAN_AUX;
> + else if (ret >= 0)
> + aux_channels[i] = ret;
> + else
> + return dev_err_probe(&client->dev, ret,
> + "Failed getting aux-channel.\n");
> + }
> +
> + ret = ltc3208_update_aux_dac(data, aux_channels[0], aux_channels[1],
> + aux_channels[2], aux_channels[3]);
Why not just aux_channels and pull the values out in the function.
> + if (ret)
> + return dev_err_probe(&client->dev, ret,
> + "error writing to aux %u channel register.\n", i);
When is 'i' not 'LTC3208_NUM_AUX_LEDS'?
> + i2c_set_clientdata(client, data);
> +
> + device_for_each_child_node_scoped(&client->dev, child) {
> + struct ltc3208_led *led;
> + struct led_init_data init_data = {};
> +
> + ret = fwnode_property_read_u32(child, "reg", &val);
> + if (ret || val >= LTC3208_NUM_LED_GRPS)
> + return dev_err_probe(&client->dev, -EINVAL,
> + "Invalid reg property for LED\n");
Why aren't we propagating the real error?
> +
> + led = &leds[val];
> + led->client = client;
> + led->channel = val;
> + led->cdev.brightness_set_blocking = ltc3208_led_set_brightness;
> + led->cdev.max_brightness = LTC3208_MAX_BRIGHTNESS_4BIT;
> + if (val == LTC3208_CHAN_MAIN || val == LTC3208_CHAN_SUB)
> + led->cdev.max_brightness = LTC3208_MAX_BRIGHTNESS_8BIT;
> +
> + init_data.fwnode = child;
> +
> + ret = devm_led_classdev_register_ext(&client->dev, &led->cdev,
> + &init_data);
> + if (ret)
> + return dev_err_probe(&client->dev, ret,
> + "Failed to register LED %u\n", val);
> + }
> +
> + data->leds = leds;
> +
> + return 0;
> +}
> +
> +static const struct of_device_id ltc3208_match_table[] = {
> + {.compatible = "adi,ltc3208"},
> + { }
> +};
> +MODULE_DEVICE_TABLE(of, ltc3208_match_table);
> +
> +static const struct i2c_device_id ltc3208_idtable[] = {
> + { "ltc3208" },
> + { }
> +};
> +MODULE_DEVICE_TABLE(i2c, ltc3208_idtable);
> +
> +static struct i2c_driver ltc3208_driver = {
> + .driver = {
> + .name = "ltc3208",
> + .of_match_table = ltc3208_match_table,
> + },
> + .id_table = ltc3208_idtable,
> + .probe = ltc3208_probe,
> +};
> +module_i2c_driver(ltc3208_driver);
> +
> +MODULE_LICENSE("GPL");
> +MODULE_AUTHOR("Jan Carlo Roleda <jancarlo.roleda@xxxxxxxxxx>");
> +MODULE_DESCRIPTION("LTC3208 LED Driver");
>
> --
> 2.43.0
>
--
Lee Jones [李琼斯]