[PATCH v3 0/4] Configure SYR82X regulator modes via device tree

[PATCH v3 0/4] Configure SYR82X regulator modes via device tree

[Victor Krawiec]

SYR82X is a regulator that supports two modes of operation:
- Auto-PFM: Allow automatic PFM during light load (default mode)
- Forced PWM

Some boards require to configure forced PWM mode from device tree to keep
the supply ripple within acceptable limits under light load conditions.

SYR82X is a clone of FAN53555 regulator and its configuration is handled by
the fcs,fan53555 driver. All clones of FAN53555 regulator share the same
two mode of operation (Auto-PFM / Forced PWM).

However the implementation in the driver is currently restricted to SYR82X
as it is the only tested hardware. 
---
V1 -> V2:
  - Add missing dt binding property description
  - Naming changes since all clones handled by the fan53555 driver share
    the same two modes
      - dt binding header name changed from silergy,syr82x.h to 
        fcs,fan53555-regulator.h - dt binding header constants changed from
        SYR82X_* to FAN53555_* 
      - fan53555_silergy_map_mode() changed to fan53555_map_mode()
  - More detailed commit descriptions

V2 -> V3:
  - Remove garbage in commit description of first commit and duplicated
    SoB line
 
Victor Krawiec (4):
  regulator: dt-bindings: Add fan53555 allowed modes
  arm64: dts: rk3399-rock-4c-plus: Use device tree bindings constants
    for SYR828 regulator initial mode
  arm64: dts: rockchip: rk3399-hugsun-x99: Update SYR828 regulator
    initial mode to use device tree bindings constants
  regulator: fan53555: Add support for mode operations on Silergy
    devices

 .../bindings/regulator/fcs,fan53555.yaml      |  8 +++++
 .../boot/dts/rockchip/rk3399-hugsun-x99.dts   |  3 +-
 .../boot/dts/rockchip/rk3399-rock-4c-plus.dts |  3 +-
 drivers/regulator/fan53555.c                  | 29 ++++++++++++++++++-
 .../regulator/fcs,fan53555-regulator.h        | 20 +++++++++++++
 5 files changed, 60 insertions(+), 3 deletions(-)
 create mode 100644 include/dt-bindings/regulator/fcs,fan53555-regulator.h


base-commit: 62b9452f5fc4d007f1ea31cfdbb53a96d91e3c66
-- 
2.43.0

[PATCH v3 1/4] regulator: dt-bindings: Add fan53555 allowed modes

[Victor Krawiec]

Fairchild FAN53555 and its clone from Rockchip, Silergy and TCS support
two modes of operation:
- Auto-PFM: Allow automatic PFM during light load (default mode)
- Forced PWM

Some boards require forced PWM mode to keep the supply ripple within
acceptable limits under light load conditions.

Regulator mode indexes are starting from 1 to keep backward compatibility
with existing device trees

Signed-off-by: Victor Krawiec <victor.krawiec@arturia.com>
---
 .../bindings/regulator/fcs,fan53555.yaml      |  8 ++++++++
 .../regulator/fcs,fan53555-regulator.h        | 20 +++++++++++++++++++
 2 files changed, 28 insertions(+)
 create mode 100644 include/dt-bindings/regulator/fcs,fan53555-regulator.h

diff --git a/Documentation/devicetree/bindings/regulator/fcs,fan53555.yaml b/Documentation/devicetree/bindings/regulator/fcs,fan53555.yaml
index 69bae90fc4b2..9a18891f721e 100644
--- a/Documentation/devicetree/bindings/regulator/fcs,fan53555.yaml
+++ b/Documentation/devicetree/bindings/regulator/fcs,fan53555.yaml
@@ -48,6 +48,12 @@ properties:
       VSEL0 register. When this pin is HIGH, VOUT is set by the VSEL1 register.
     maxItems: 1
 
+  regulator-initial-mode:
+    enum:
+      [ 1, 2 ]
+    description:
+      Defined in include/dt-bindings/regulator/fcs,fan53555-regulator.h
+
 required:
   - compatible
   - reg
@@ -56,6 +62,7 @@ unevaluatedProperties: false
 
 examples:
   - |
+    #include <dt-bindings/regulator/fcs,fan53555-regulator.h>
     i2c {
         #address-cells = <1>;
         #size-cells = <0>;
@@ -68,6 +75,7 @@ examples:
             regulator-max-microvolt = <1800000>;
             vin-supply = <&parent_reg>;
             fcs,suspend-voltage-selector = <1>;
+            regulator-initial-mode = <FAN53555_REGULATOR_MODE_FORCE_PWM>;
         };
     };
 ...
diff --git a/include/dt-bindings/regulator/fcs,fan53555-regulator.h b/include/dt-bindings/regulator/fcs,fan53555-regulator.h
new file mode 100644
index 000000000000..ce3d6c63a0f1
--- /dev/null
+++ b/include/dt-bindings/regulator/fcs,fan53555-regulator.h
@@ -0,0 +1,20 @@
+/* SPDX-License-Identifier: (GPL-2.0-only OR BSD-2-Clause) */
+/*
+ * Copyright (c) 2026 Arturia - All rights reserved.
+ *
+ * Device Tree binding constants for the FAN53555 PMIC regulator
+ */
+
+#ifndef _DT_BINDINGS_REGULATOR_FAN53555_H
+#define _DT_BINDINGS_REGULATOR_FAN53555_H
+
+/*
+ * Constants to specify regulator modes in device tree for SYR82X regulators
+ * FAN53555_REGULATOR_MODE_FORCE_PWM:	Force fixed PWM mode
+ * FAN53555_REGULATOR_MODE_AUTO:	Allow auto-PFM mode during light load
+ */
+
+#define FAN53555_REGULATOR_MODE_FORCE_PWM	1
+#define FAN53555_REGULATOR_MODE_AUTO		2
+
+#endif

base-commit: 62b9452f5fc4d007f1ea31cfdbb53a96d91e3c66
-- 
2.43.0

[PATCH v3 2/4] arm64: dts: rk3399-rock-4c-plus: Use device tree bindings constants for SYR828 regulator initial mode

[Victor Krawiec]

Prefer using a device tree bindings constant over a hardcoded value

Signed-off-by: Victor Krawiec <victor.krawiec@arturia.com>
---
 arch/arm64/boot/dts/rockchip/rk3399-rock-4c-plus.dts | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/arch/arm64/boot/dts/rockchip/rk3399-rock-4c-plus.dts b/arch/arm64/boot/dts/rockchip/rk3399-rock-4c-plus.dts
index d9ff777b4913..0387172cb470 100644
--- a/arch/arm64/boot/dts/rockchip/rk3399-rock-4c-plus.dts
+++ b/arch/arm64/boot/dts/rockchip/rk3399-rock-4c-plus.dts
@@ -7,6 +7,7 @@
 
 /dts-v1/;
 #include <dt-bindings/leds/common.h>
+#include <dt-bindings/regulator/fcs,fan53555-regulator.h>
 #include "rk3399-t.dtsi"
 
 / {
@@ -448,7 +449,7 @@ vdd_gpu: regulator@41 {
 		regulator-always-on;
 		regulator-boot-on;
 		vin-supply = <&vcc5v0_sys>;
-		regulator-initial-mode = <1>; /* 1:force PWM 2:auto */
+		regulator-initial-mode = <FAN53555_REGULATOR_MODE_FORCE_PWM>;
 		regulator-state-mem {
 			regulator-off-in-suspend;
 		};
-- 
2.43.0

[PATCH v3 3/4] arm64: dts: rockchip: rk3399-hugsun-x99: Update SYR828 regulator initial mode to use device tree bindings constants

[Victor Krawiec]

Prefer using a device tree bindings constant over a hardcoded value

Signed-off-by: Victor Krawiec <victor.krawiec@arturia.com>
---
 arch/arm64/boot/dts/rockchip/rk3399-hugsun-x99.dts | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/arch/arm64/boot/dts/rockchip/rk3399-hugsun-x99.dts b/arch/arm64/boot/dts/rockchip/rk3399-hugsun-x99.dts
index 352c8efb37e0..a2b7bab63553 100644
--- a/arch/arm64/boot/dts/rockchip/rk3399-hugsun-x99.dts
+++ b/arch/arm64/boot/dts/rockchip/rk3399-hugsun-x99.dts
@@ -3,6 +3,7 @@
 #include <dt-bindings/pwm/pwm.h>
 #include <dt-bindings/input/input.h>
 #include <dt-bindings/interrupt-controller/irq.h>
+#include <dt-bindings/regulator/fcs,fan53555-regulator.h>
 #include "rk3399.dtsi"
 
 / {
@@ -237,7 +238,7 @@ vdd_gpu: syr828@41 {
 		regulator-always-on;
 		regulator-boot-on;
 		vin-supply = <&vcc_sys>;
-		regulator-initial-mode = <1>;
+		regulator-initial-mode = <FAN53555_REGULATOR_MODE_FORCE_PWM>;
 		regulator-state-mem {
 			regulator-off-in-suspend;
 		};
-- 
2.43.0

[PATCH v3 4/4] regulator: fan53555: Add support for mode operations on Silergy devices

[Victor Krawiec]

Make the PWM mode configurable from devicetree. Some boards require forced
PWM mode to keep the supply ripple within acceptable limits under light
load conditions

Support is restricted to Silergy manufacturer as it is the only one
currently tested.

Signed-off-by: Victor Krawiec <victor.krawiec@arturia.com>
---
 drivers/regulator/fan53555.c | 29 ++++++++++++++++++++++++++++-
 1 file changed, 28 insertions(+), 1 deletion(-)

diff --git a/drivers/regulator/fan53555.c b/drivers/regulator/fan53555.c
index c282236959b1..85f46874eb49 100644
--- a/drivers/regulator/fan53555.c
+++ b/drivers/regulator/fan53555.c
@@ -21,6 +21,7 @@
 #include <linux/regulator/machine.h>
 #include <linux/regulator/of_regulator.h>
 #include <linux/slab.h>
+#include <dt-bindings/regulator/fcs,fan53555-regulator.h>
 
 /* Voltage setting */
 #define FAN53555_VSEL0		0x00
@@ -389,6 +390,18 @@ static int rk8602_voltages_setup_rockchip(struct fan53555_device_info *di)
 	return 0;
 }
 
+static inline unsigned int fan53555_map_mode(unsigned int mode)
+{
+	switch (mode) {
+	case FAN53555_REGULATOR_MODE_FORCE_PWM:
+		return REGULATOR_MODE_FAST;
+	case FAN53555_REGULATOR_MODE_AUTO:
+		return REGULATOR_MODE_NORMAL;
+	default:
+		return REGULATOR_MODE_INVALID;
+	}
+}
+
 static int fan53555_voltages_setup_silergy(struct fan53555_device_info *di)
 {
 	/* Init voltage range and step */
@@ -586,6 +599,17 @@ static int fan53555_device_setup(struct fan53555_device_info *di,
 	return ret;
 }
 
+static void fan53555_device_mode_map_setup(struct fan53555_device_info *di)
+{
+	switch (di->vendor) {
+	case FAN53555_VENDOR_SILERGY:
+		di->desc.of_map_mode = fan53555_map_mode;
+		break;
+	default:
+		break;
+	}
+}
+
 static int fan53555_regulator_register(struct fan53555_device_info *di,
 			struct regulator_config *config)
 {
@@ -686,6 +710,10 @@ static int fan53555_regulator_probe(struct i2c_client *client)
 	if (!di)
 		return -ENOMEM;
 
+	di->vendor = (uintptr_t)i2c_get_match_data(client);
+
+	fan53555_device_mode_map_setup(di);
+
 	pdata = dev_get_platdata(&client->dev);
 	if (!pdata)
 		pdata = fan53555_parse_dt(&client->dev, np, &di->desc);
@@ -695,7 +723,6 @@ static int fan53555_regulator_probe(struct i2c_client *client)
 				     "Platform data not found!\n");
 
 	di->regulator = pdata->regulator;
-	di->vendor = (uintptr_t)i2c_get_match_data(client);
 	if (!dev_fwnode(&client->dev)) {
 		/* if no ramp constraint set, get the pdata ramp_delay */
 		if (!di->regulator->constraints.ramp_delay) {
-- 
2.43.0

Re: [PATCH v3 4/4] regulator: fan53555: Add support for mode operations on Silergy devices

[sashiko-bot]

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:

Pre-existing issues:
- [High] The function `fan53555_set_mode` unconditionally returns 0 and ignores the error return value from `regmap_update_bits`, falsely informing the regulator core that mode changes were successful even when underlying I2C transfers fail.
--

--- Patch [4]: [PATCH v3 4/4] regulator: fan53555: Add support for mode operations on Silergy devices ---
commit 0810cb3eb74c80e6af381ca179af7f69456b49a8
Author: Victor Krawiec <victor.krawiec@arturia.com>

    regulator: fan53555: Add support for mode operations on Silergy devices
    
    Make the PWM mode configurable from devicetree. Some boards require forced
    PWM mode to keep the supply ripple within acceptable limits under light
    load conditions
    
    Support is restricted to Silergy manufacturer as it is the only one
    currently tested.

> diff --git a/drivers/regulator/fan53555.c b/drivers/regulator/fan53555.c
> index c282236959b18..85f46874eb49d 100644
> --- a/drivers/regulator/fan53555.c
> +++ b/drivers/regulator/fan53555.c

[ ... ]

> @@ -389,6 +390,18 @@ static int rk8602_voltages_setup_rockchip(struct fan53555_device_info *di)
>  	return 0;
>  }
>  
> +static inline unsigned int fan53555_map_mode(unsigned int mode)
> +{
> +	switch (mode) {
> +	case FAN53555_REGULATOR_MODE_FORCE_PWM:
> +		return REGULATOR_MODE_FAST;
> +	case FAN53555_REGULATOR_MODE_AUTO:
> +		return REGULATOR_MODE_NORMAL;
> +	default:
> +		return REGULATOR_MODE_INVALID;
> +	}
> +}
> +

[Severity: High]
This is a pre-existing issue, but since this patch enables setting the
mode from the device tree via fan53555_map_mode(), should fan53555_set_mode()
check the return value of regmap_update_bits()?

Looking at fan53555_set_mode() in drivers/regulator/fan53555.c, it
unconditionally returns 0 and ignores the error return value from
regmap_update_bits():

drivers/regulator/fan53555.c:fan53555_set_mode() {
    ...
	switch (mode) {
	case REGULATOR_MODE_FAST:
		regmap_update_bits(rdev->regmap, di->mode_reg,
				   di->mode_mask, di->mode_mask);
		break;
	case REGULATOR_MODE_NORMAL:
		regmap_update_bits(rdev->regmap, di->vol_reg, di->mode_mask, 0);
		break;
	default:
		return -EINVAL;
	}
	return 0;
}

If an I2C transfer fails here, the regulator hardware could remain in
auto mode while the kernel core falsely believes it was successfully set to
forced PWM mode. Could this lead to the unacceptable voltage ripple under
light load conditions mentioned in the commit message?

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260723094001.120264-1-victor.krawiec@arturia.com?part=4

Re: [PATCH v3 1/4] regulator: dt-bindings: Add fan53555 allowed modes

[sashiko-bot]

Thank you for your contribution! Sashiko AI review found 3 potential issue(s) to consider:

New issues:
- [Medium] The device tree binding introduces the `regulator-initial-mode` property, but the corresponding driver lacks the `.of_map_mode` callback required to use it.
- [Low] The patch defines constraints for `regulator-initial-mode` but fails to define constraints for `regulator-allowed-modes`.

Pre-existing issues:
- [High] A preexisting bug in `fan53555_set_mode()` corrupts the voltage register when setting `REGULATOR_MODE_NORMAL` on certain vendor variants like RK8602.
--

--- Patch [1]: [PATCH v3 1/4] regulator: dt-bindings: Add fan53555 allowed modes ---
commit 23beca6378b28eed5bc10d3625b311119ecc5eb7
Author: Victor Krawiec <victor.krawiec@arturia.com>

regulator: dt-bindings: Add fan53555 allowed modes

This commit adds device tree binding constants and schema properties to
specify the initial operation mode (Auto-PFM or Forced PWM) for the
FAN53555 PMIC regulator.

> diff --git a/Documentation/devicetree/bindings/regulator/fcs,fan53555.yaml b/Documentation/devicetree/bindings/regulator/fcs,fan53555.yaml
> index 69bae90fc4b28..9a18891f721e6 100644
> --- a/Documentation/devicetree/bindings/regulator/fcs,fan53555.yaml
> +++ b/Documentation/devicetree/bindings/regulator/fcs,fan53555.yaml
> @@ -48,6 +48,12 @@ properties:
>        VSEL0 register. When this pin is HIGH, VOUT is set by the VSEL1 register.
>      maxItems: 1
>  
> +  regulator-initial-mode:
> +    enum:
> +      [ 1, 2 ]
> +    description:
> +      Defined in include/dt-bindings/regulator/fcs,fan53555-regulator.h
> +

[Severity: Low]
The commit message mentions adding allowed modes, but the schema only defines
constraints for regulator-initial-mode. Does this leave
regulator-allowed-modes unconstrained, allowing device trees to pass
dt_binding_check even if they specify invalid integers like
regulator-allowed-modes = <99>?

[Severity: Medium]
Will the regulator core actually apply this initial mode?

The fan53555 driver currently lacks the .of_map_mode callback:

drivers/regulator/fan53555.c:fan53555_regulator_register() {
    ...
    rdesc->vsel_mask = BIT(fls(di->vsel_count - 1)) - 1;
    rdesc->owner = THIS_MODULE;

    rdev = devm_regulator_register(di->dev, &di->desc, config);
    ...
}

Without an .of_map_mode implementation to translate the DT integer into a
standard REGULATOR_MODE_* constant, won't the regulator core ignore this
property completely and fail to set the mode?

[ ... ]

> diff --git a/include/dt-bindings/regulator/fcs,fan53555-regulator.h b/include/dt-bindings/regulator/fcs,fan53555-regulator.h
> new file mode 100644
> index 0000000000000..ce3d6c63a0f1f
> --- /dev/null
> +++ b/include/dt-bindings/regulator/fcs,fan53555-regulator.h
> @@ -0,0 +1,20 @@
> +/* SPDX-License-Identifier: (GPL-2.0-only OR BSD-2-Clause) */
> +/*
> + * Copyright (c) 2026 Arturia - All rights reserved.
> + *
> + * Device Tree binding constants for the FAN53555 PMIC regulator
> + */
> +
> +#ifndef _DT_BINDINGS_REGULATOR_FAN53555_H
> +#define _DT_BINDINGS_REGULATOR_FAN53555_H
> +
> +/*
> + * Constants to specify regulator modes in device tree for SYR82X regulators
> + * FAN53555_REGULATOR_MODE_FORCE_PWM:	Force fixed PWM mode
> + * FAN53555_REGULATOR_MODE_AUTO:	Allow auto-PFM mode during light load
> + */
> +
> +#define FAN53555_REGULATOR_MODE_FORCE_PWM	1
> +#define FAN53555_REGULATOR_MODE_AUTO		2

[Severity: High]
This is a pre-existing issue, but I noticed a potential hardware register
corruption in the driver when setting modes.

In the fan53555 driver, when setting REGULATOR_MODE_NORMAL, does the code
mistakenly clear the mode mask from the voltage register instead of the mode
register?

drivers/regulator/fan53555.c:fan53555_set_mode() {
    ...
    case REGULATOR_MODE_NORMAL:
        regmap_update_bits(rdev->regmap, di->vol_reg, di->mode_mask, 0);
        break;
    ...
}

For variants like the RK8602 where mode_reg and vol_reg are distinct
hardware registers, clearing di->mode_mask (e.g. BIT(6)) in di->vol_reg
could silently drop the output voltage by 64 steps (400mV). Could this lead
to severe undervolting and system crashes if a consumer attempts to change
the mode?

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260723094001.120264-1-victor.krawiec@arturia.com?part=1