[PATCH v4 0/6] hwmon: (lm63) Add Sensylink CTF2301 support

[PATCH v4 0/6] hwmon: (lm63) Add Sensylink CTF2301 support

[Troy Mitchell]

Sensylink CTF2301 is a system-level thermal management solution chip.

The CTF2301 is an I2C/SMBus compatible device featuring:
  - One local temperature sensor with +/-0.5 degree C accuracy and
    0.0625 degree C resolution.
  - One remote temperature sensor for external diode-connected
    transistors, offering +/-1 degree C accuracy and 0.125 degree C
    resolution (temperature range: -40 degree C to +125 degree C).
  - An integrated PWM fan controller capable of operating in two modes:
      1. Direct-DCY: Open-loop direct duty cycle control.
      2. Auto-Temp: Closed-loop automatic fan speed control based on
         measured temperature.
  - A 1-channel fan speed monitor (TACH input) for RPM measurement.

As Guenter pointed out on v1, the CTF2301 register map is a near match
for the LM63/LM96163 family (local and remote temperature, tachometer and
PWM fan control), so this version drops the standalone ctf2301 driver
and instead adds CTF2301 as a new chip type in the existing lm63 driver.
This reuses the shared temperature, limit, alarm and PWM handling instead
of duplicating it.

The CTF2301-specific differences are handled on top of lm63: 12-bit local
temperature input and high-limit values, nine conversion rates, 12
fan-control lookup table entries, and optional 22.5 kHz high-resolution
PWM. When no IRQ is described, the shared ALERT/TACH pin is configured as
a tachometer input. The LM63 family can also register as a thermal cooling
device when the device tree provides the #cooling-cells property.

Tested on a SpacemiT K3 CoM260 by booting an initramfs and verifying the
CTF2301 probe. Temperature, tachometer, PWM, and conversion-rate behavior
were checked against direct I2C register accesses and the datasheet.

Datasheet:
https://www.sensylink.com/upload/1/net.sensylink.portal/1689557281035.pdf

Register description:
https://github.com/TroyMitchell911/ctf2301-datasheet

Signed-off-by: Troy Mitchell <troy.mitchell@linux.dev>
---
Changes in v4:
- Rebase onto v7.3-rc1, which includes the SpacemiT I2C timing
  prerequisite, and remove the obsolete dependency note.
- Do not add the SENSORS_LM63 THERMAL dependency suggested on v3 because
  CONFIG_THERMAL is boolean and v7.3-rc1 removes such tautological
  expressions.
- Preserve the upstream five-bit PWM frequency multiplier mask in the
  LM63 initialization error-handling path.
- Sort the chip type enum and remove redundant CTF2301 8-bit local
  temperature cache updates.
- Serialize CTF2301 12-bit local temperature cache reads.
- Keep PWM configuration caches synchronized after partial I2C write
  failures.
- Override the CoM260 I2C6 pin configuration through &i2c6_1_cfg,
  as suggested by Yixun.
- Add Krzysztof's Acked-by and Reviewed-by tags.
- Link to v3: https://patch.msgid.link/20260805-ctl2301-v3-0-fd4b8ffb5b0d@linux.dev

Changes in v3:
- Add the Sensylink website link to the vendor-prefix commit.
- Preserve the optional interrupt from the trivial-devices binding.
- Explain in the commit message of patch 3/6 why CTF2301 cannot use
  national,lm63 as a fallback compatible.
- Check both bytes of CTF2301 local temperature values before caching them.
- Use the schema-valid 25 mA drive strength for the K3 I2C6 pins.
- Link to v2: https://patch.msgid.link/20260721-ctl2301-v2-0-ef5fb342872a@linux.dev

Changes in v2:
- Replace the standalone CTF2301 driver with a new lm63 chip type, as
  suggested by Guenter, reusing the existing temperature, limit, alarm,
  tachometer, PWM, and automatic fan-control paths.
- Add the temperature thresholds, alarms, and automatic fan-control lookup
  table support which were missing from the v1 driver.
- Handle CTF2301-specific 12-bit local temperature values, nine conversion
  rates, 12 lookup table entries, shared ALERT/TACH pin, and optional
  22.5 kHz high-resolution PWM.
- Propagate SMBus failures from lm63 initialization and write paths.
- Add optional thermal cooling-device support for the LM63 family using
  the existing manual PWM control path.
- Move the National LM63-family compatibles from trivial-devices.yaml to a
  dedicated binding which permits #cooling-cells.
- Add CTF2301 to the LM63-family binding instead of keeping the standalone
  schema from v1. Do not declare #cooling-cells on K3 CoM260 yet because
  the board has no thermal consumer.
- Fix the binding errors reported by Rob by using a plain compatible string
  and wrapping the description.
- Document CTF2301 support in the lm63 hwmon documentation and provide the
  detailed register description used for review.
- Add the K3 CoM260 I2C6 pinctrl state and CTF2301 device node for hardware
  testing.
- Link to v1: https://lore.kernel.org/r/20250916-ctl2301-v1-0-97e7c84f2c47@linux.dev

To: Rob Herring <robh@kernel.org>
To: Krzysztof Kozlowski <krzk+dt@kernel.org>
To: Conor Dooley <conor+dt@kernel.org>
To: Guenter Roeck <linux@roeck-us.net>
To: Paul Walmsley <pjw@kernel.org>
To: Palmer Dabbelt <palmer@dabbelt.com>
To: Albert Ou <aou@eecs.berkeley.edu>
To: Alexandre Ghiti <alex@ghiti.fr>
To: Yixun Lan <dlan@kernel.org>
To: Jonathan Corbet <corbet@lwn.net>
To: Shuah Khan <skhan@linuxfoundation.org>
To: Randy Dunlap <rdunlap@infradead.org>
Cc: devicetree@vger.kernel.org
Cc: linux-kernel@vger.kernel.org
Cc: linux-hwmon@vger.kernel.org
Cc: linux-riscv@lists.infradead.org
Cc: spacemit@lists.linux.dev
Cc: linux-doc@vger.kernel.org

---
Troy Mitchell (6):
      dt-bindings: vendor-prefixes: Add Sensylink
      dt-bindings: hwmon: Move LM63 family to a dedicated binding
      dt-bindings: hwmon: Add Sensylink CTF2301
      hwmon: (lm63) Add Sensylink CTF2301 support
      riscv: dts: spacemit: Add K3 I2C6 pinctrl state
      riscv: dts: spacemit: Add CTF2301 on K3 CoM260

 .../devicetree/bindings/hwmon/national,lm63.yaml   |  55 +++
 .../devicetree/bindings/trivial-devices.yaml       |   6 -
 .../devicetree/bindings/vendor-prefixes.yaml       |   2 +
 Documentation/hwmon/lm63.rst                       |  22 ++
 arch/riscv/boot/dts/spacemit/k3-com260.dtsi        |  17 +
 arch/riscv/boot/dts/spacemit/k3-pinctrl.dtsi       |  11 +
 drivers/hwmon/Kconfig                              |   8 +-
 drivers/hwmon/lm63.c                               | 426 +++++++++++++++++----
 8 files changed, 462 insertions(+), 85 deletions(-)
---
base-commit: cee9395acd8043be0644b25c34bfa86623f2b935
change-id: 20250916-ctl2301-0416b073c280

Best regards,
--  
Troy Mitchell <troy.mitchell@linux.dev>

[PATCH v4 1/6] dt-bindings: vendor-prefixes: Add Sensylink

[Troy Mitchell]

Add the vendor prefix for Sensylink Microelectronics.

Link: https://en.sensylink.com/
Acked-by: Krzysztof Kozlowski <krzysztof.kozlowski@oss.qualcomm.com>
Signed-off-by: Troy Mitchell <troy.mitchell@linux.dev>
---
 Documentation/devicetree/bindings/vendor-prefixes.yaml | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/Documentation/devicetree/bindings/vendor-prefixes.yaml b/Documentation/devicetree/bindings/vendor-prefixes.yaml
index ba2002969373..4a5cf436eb14 100644
--- a/Documentation/devicetree/bindings/vendor-prefixes.yaml
+++ b/Documentation/devicetree/bindings/vendor-prefixes.yaml
@@ -1495,6 +1495,8 @@ patternProperties:
     description: Sensirion AG
   "^sensortek,.*":
     description: Sensortek Technology Corporation
+  "^sensylink,.*":
+    description: Sensylink Microelectronics Technology Co., Ltd.
   "^sercomm,.*":
     description: Sercomm (Suzhou) Corporation
   "^sff,.*":

-- 
2.55.0

Re: [PATCH v4 1/6] dt-bindings: vendor-prefixes: Add Sensylink

[sashiko-bot]

> Add the vendor prefix for Sensylink Microelectronics.
> 
> Link: https://en.sensylink.com/
> Acked-by: Krzysztof Kozlowski <krzysztof.kozlowski@oss.qualcomm.com>
> Signed-off-by: Troy Mitchell <troy.mitchell@linux.dev>

Sashiko has reviewed this patch and found no issues. It looks great!

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260831-ctl2301-v4-0-b335d52a0570@linux.dev?part=1

[PATCH v4 2/6] dt-bindings: hwmon: Move LM63 family to a dedicated binding

[Troy Mitchell]

Move the National Semiconductor LM63, LM64, and LM96163 compatibles out
of trivial-devices.yaml and into a dedicated binding. Describe their
integrated fan control and preserve the optional interrupt used by the
ALERT output. Permit use as thermal cooling devices via #cooling-cells.

Reviewed-by: Krzysztof Kozlowski <krzysztof.kozlowski@oss.qualcomm.com>
Signed-off-by: Troy Mitchell <troy.mitchell@linux.dev>
---
 .../devicetree/bindings/hwmon/national,lm63.yaml   | 53 ++++++++++++++++++++++
 .../devicetree/bindings/trivial-devices.yaml       |  6 ---
 2 files changed, 53 insertions(+), 6 deletions(-)

diff --git a/Documentation/devicetree/bindings/hwmon/national,lm63.yaml b/Documentation/devicetree/bindings/hwmon/national,lm63.yaml
new file mode 100644
index 000000000000..f2e0993cfbef
--- /dev/null
+++ b/Documentation/devicetree/bindings/hwmon/national,lm63.yaml
@@ -0,0 +1,53 @@
+# SPDX-License-Identifier: GPL-2.0-only OR BSD-2-Clause
+%YAML 1.2
+---
+$id: http://devicetree.org/schemas/hwmon/national,lm63.yaml#
+$schema: http://devicetree.org/meta-schemas/core.yaml#
+
+title: LM63-compatible temperature sensors with integrated fan control
+
+maintainers:
+  - Guenter Roeck <linux@roeck-us.net>
+
+description: |
+  The LM63 family provides local and remote temperature monitoring together
+  with fan speed monitoring and PWM fan control. Supported devices include
+  the National Semiconductor LM63, LM64 and LM96163.
+
+allOf:
+  - $ref: hwmon-common.yaml#
+
+properties:
+  compatible:
+    enum:
+      - national,lm63
+      - national,lm64
+      - national,lm96163
+
+  reg:
+    maxItems: 1
+
+  interrupts:
+    maxItems: 1
+
+  "#cooling-cells":
+    const: 2
+
+required:
+  - compatible
+  - reg
+
+unevaluatedProperties: false
+
+examples:
+  - |
+    i2c {
+        #address-cells = <1>;
+        #size-cells = <0>;
+
+        temperature-sensor@4c {
+            compatible = "national,lm63";
+            reg = <0x4c>;
+            #cooling-cells = <2>;
+        };
+    };
diff --git a/Documentation/devicetree/bindings/trivial-devices.yaml b/Documentation/devicetree/bindings/trivial-devices.yaml
index 2de8eb09cb7d..e46912ba2fd6 100644
--- a/Documentation/devicetree/bindings/trivial-devices.yaml
+++ b/Documentation/devicetree/bindings/trivial-devices.yaml
@@ -368,16 +368,10 @@ properties:
           - mps,mpq82d00
             # Murata D1U74T-W power supply unit
           - murata,d1u74t
-            # Temperature sensor with integrated fan control
-          - national,lm63
-            # Temperature sensor with integrated fan control
-          - national,lm64
             # Temperature sensor
           - national,lm95235
             # Temperature sensor
           - national,lm95245
-            # Temperature sensor with integrated fan control
-          - national,lm96163
             # Serial Interface ACPI-Compatible Microprocessor System Hardware Monitor
           - national,lm80
             # Serial Interface ACPI-Compatible Microprocessor System Hardware Monitor

-- 
2.55.0

Re: [PATCH v4 2/6] dt-bindings: hwmon: Move LM63 family to a dedicated binding

[sashiko-bot]

> Move the National Semiconductor LM63, LM64, and LM96163 compatibles out
> of trivial-devices.yaml and into a dedicated binding. Describe their
> integrated fan control and preserve the optional interrupt used by the
> ALERT output. Permit use as thermal cooling devices via #cooling-cells.
> 
> Reviewed-by: Krzysztof Kozlowski <krzysztof.kozlowski@oss.qualcomm.com>
> Signed-off-by: Troy Mitchell <troy.mitchell@linux.dev>

Sashiko has reviewed this patch and found no issues. It looks great!

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260831-ctl2301-v4-0-b335d52a0570@linux.dev?part=2

[PATCH v4 3/6] dt-bindings: hwmon: Add Sensylink CTF2301

[Troy Mitchell]

Add the Sensylink CTF2301 temperature sensor and fan controller to the
LM63 family binding. The chip follows the LM63 register model and can
provide a thermal cooling device when #cooling-cells is present.

Do not provide a national,lm63 fallback because the CTF2301 is not fully
compatible with LM63. It uses different conversion-rate encodings,
extended local-temperature registers, additional fan lookup table
entries, and device-specific PWM and ALERT/TACH configuration. An older
LM63 driver would therefore not operate the device correctly.

Acked-by: Krzysztof Kozlowski <krzysztof.kozlowski@oss.qualcomm.com>
Signed-off-by: Troy Mitchell <troy.mitchell@linux.dev>
---
 Documentation/devicetree/bindings/hwmon/national,lm63.yaml | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

diff --git a/Documentation/devicetree/bindings/hwmon/national,lm63.yaml b/Documentation/devicetree/bindings/hwmon/national,lm63.yaml
index f2e0993cfbef..4c8d0a15fbd1 100644
--- a/Documentation/devicetree/bindings/hwmon/national,lm63.yaml
+++ b/Documentation/devicetree/bindings/hwmon/national,lm63.yaml
@@ -12,7 +12,8 @@ maintainers:
 description: |
   The LM63 family provides local and remote temperature monitoring together
   with fan speed monitoring and PWM fan control. Supported devices include
-  the National Semiconductor LM63, LM64 and LM96163.
+  the National Semiconductor LM63, LM64 and LM96163, and the Sensylink
+  CTF2301.
 
 allOf:
   - $ref: hwmon-common.yaml#
@@ -23,6 +24,7 @@ properties:
       - national,lm63
       - national,lm64
       - national,lm96163
+      - sensylink,ctf2301
 
   reg:
     maxItems: 1

-- 
2.55.0

Re: [PATCH v4 3/6] dt-bindings: hwmon: Add Sensylink CTF2301

[sashiko-bot]

> Add the Sensylink CTF2301 temperature sensor and fan controller to the
> LM63 family binding. The chip follows the LM63 register model and can
> provide a thermal cooling device when #cooling-cells is present.
> 
> Do not provide a national,lm63 fallback because the CTF2301 is not fully
> compatible with LM63. It uses different conversion-rate encodings,
> extended local-temperature registers, additional fan lookup table
> entries, and device-specific PWM and ALERT/TACH configuration. An older
> LM63 driver would therefore not operate the device correctly.
> 
> Acked-by: Krzysztof Kozlowski <krzysztof.kozlowski@oss.qualcomm.com>
> Signed-off-by: Troy Mitchell <troy.mitchell@linux.dev>

Sashiko has reviewed this patch and found no issues. It looks great!

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260831-ctl2301-v4-0-b335d52a0570@linux.dev?part=3

[PATCH v4 4/6] hwmon: (lm63) Add Sensylink CTF2301 support

[Troy Mitchell]

Add Sensylink CTF2301 support to the LM63 driver. CTF2301 follows
the LM63-style temperature, tachometer, and PWM register layout,
while using a 12-bit local temperature register and requiring the
ALERT/TACH pin to be configured for tachometer input.

Register an optional thermal cooling device when the device tree
node provides #cooling-cells, allowing thermal zones to control
the fan through the existing PWM path without changing behavior
for existing LM63-family users.

Signed-off-by: Troy Mitchell <troy.mitchell@linux.dev>
---
 Documentation/hwmon/lm63.rst |  22 +++
 drivers/hwmon/Kconfig        |   8 +-
 drivers/hwmon/lm63.c         | 426 +++++++++++++++++++++++++++++++++++--------
 3 files changed, 377 insertions(+), 79 deletions(-)

diff --git a/Documentation/hwmon/lm63.rst b/Documentation/hwmon/lm63.rst
index 9e27367d7405..c40bd0dfe591 100644
--- a/Documentation/hwmon/lm63.rst
+++ b/Documentation/hwmon/lm63.rst
@@ -33,6 +33,16 @@ Supported chips:
 
 	       http://www.national.com/pf/LM/LM96163.html
 
+  * Sensylink CTF2301
+
+    Prefix: 'ctf2301'
+
+    Addresses scanned: none
+
+    Datasheet: https://www.sensylink.com/upload/1/net.sensylink.portal/1689557281035.pdf
+
+    Register description: https://github.com/TroyMitchell911/ctf2301-datasheet
+
 
 Author: Jean Delvare <jdelvare@suse.de>
 
@@ -62,6 +72,8 @@ value have to be masked out. The value is still 16 bit in width.
 
 All temperature values are given in degrees Celsius. Resolution is 1.0
 degree for the local temperature, 0.125 degree for the remote temperature.
+The CTF2301 local temperature input and limit have a resolution of 0.0625
+degree.
 
 The fan speed is measured using a tachometer. Contrary to most chips which
 store the value in an 8-bit register and have a selectable clock divider
@@ -93,3 +105,13 @@ support these GPIO lines at present.
 The LM96163 is an enhanced version of LM63 with improved temperature accuracy
 and better PWM resolution. For LM96163, the external temperature sensor type is
 configurable as CPU embedded diode(1) or 3904 transistor(2).
+
+The CTF2301 is register-compatible with the LM63 family and provides 12 fan
+control lookup table entries. It supports 8-bit PWM resolution when configured
+for a 22.5 kHz PWM frequency.
+
+Device tree nodes may declare ``#cooling-cells`` to register the fan controller
+with the thermal framework. The cooling state is mapped to the existing
+``pwm1`` range from 0 to 255. The driver keeps the current automatic or manual
+fan control mode during probe and switches to manual control when the thermal
+framework first requests a cooling state.
diff --git a/drivers/hwmon/Kconfig b/drivers/hwmon/Kconfig
index fecff8610ea8..5d87ad3e7a20 100644
--- a/drivers/hwmon/Kconfig
+++ b/drivers/hwmon/Kconfig
@@ -1521,10 +1521,10 @@ config SENSORS_LM63
 	depends on I2C
 	help
 	  If you say yes here you get support for the National
-	  Semiconductor LM63, LM64, and LM96163 remote diode digital temperature
-	  sensors with integrated fan control.  Such chips are found
-	  on the Tyan S4882 (Thunder K8QS Pro) motherboard, among
-	  others.
+	  Semiconductor LM63, LM64, LM96163, and Sensylink CTF2301
+	  remote diode digital temperature sensors with integrated fan
+	  control. Such chips are found on the Tyan S4882 (Thunder K8QS
+	  Pro) motherboard, among others.
 
 	  This driver can also be built as a module. If so, the module
 	  will be called lm63.
diff --git a/drivers/hwmon/lm63.c b/drivers/hwmon/lm63.c
index da09770c05e7..5a94895242ed 100644
--- a/drivers/hwmon/lm63.c
+++ b/drivers/hwmon/lm63.c
@@ -35,13 +35,14 @@
 #include <linux/mutex.h>
 #include <linux/of.h>
 #include <linux/sysfs.h>
+#include <linux/thermal.h>
 #include <linux/types.h>
 
 /*
  * Addresses to scan
  * Address is fully defined internally and cannot be changed except for
  * LM64 which has one pin dedicated to address selection.
- * LM63 and LM96163 have address 0x4c.
+ * LM63, LM96163, and CTF2301 have address 0x4c.
  * LM64 can have address 0x18 or 0x4e.
  */
 
@@ -70,6 +71,9 @@ static const unsigned short normal_i2c[] = { 0x18, 0x4c, 0x4e, I2C_CLIENT_END };
 #define LM63_REG_LOCAL_TEMP		0x00
 #define LM63_REG_LOCAL_HIGH		0x05
 
+#define CTF2301_REG_LOCAL_TEMP_LSB	0x15
+#define CTF2301_REG_LOCAL_HIGH_LSB	0x06
+
 #define LM63_REG_REMOTE_TEMP_MSB	0x01
 #define LM63_REG_REMOTE_TEMP_LSB	0x10
 #define LM63_REG_REMOTE_OFFSET_MSB	0x11
@@ -94,6 +98,7 @@ static const unsigned short normal_i2c[] = { 0x18, 0x4c, 0x4e, I2C_CLIENT_END };
 
 #define LM63_PWM_BASE_FAST_HZ		180000
 #define LM63_PWM_BASE_SLOW_HZ		700
+#define LM63_MAX_PWM			255
 
 #define LM63_MAX_CONVRATE		9
 
@@ -121,6 +126,9 @@ static const unsigned short normal_i2c[] = { 0x18, 0x4c, 0x4e, I2C_CLIENT_END };
 							    127000), 1000)
 #define TEMP8U_TO_REG(val)	DIV_ROUND_CLOSEST(clamp_val((val), 0, \
 							    255000), 1000)
+#define TEMP12_FROM_REG(reg)	DIV_ROUND_CLOSEST((s16)(reg) * 1000, 256)
+#define TEMP12_TO_REG(val)	DIV_ROUND_CLOSEST(clamp_val((val), -128000, \
+							     127937) * 256, 1000)
 #define TEMP11_FROM_REG(reg)	((reg) / 32 * 125)
 #define TEMP11_TO_REG(val)	(DIV_ROUND_CLOSEST(clamp_val((val), -128000, \
 							     127875), 125) * 32)
@@ -132,7 +140,11 @@ static const unsigned short normal_i2c[] = { 0x18, 0x4c, 0x4e, I2C_CLIENT_END };
 #define UPDATE_INTERVAL(max, rate) \
 			((1000 << (LM63_MAX_CONVRATE - (rate))) / (max))
 
-enum chips { lm63, lm64, lm96163 };
+enum chips { ctf2301, lm63, lm64, lm96163 };
+
+static const unsigned int ctf2301_update_intervals[] = {
+	20000, 10000, 4902, 2463, 1000, 615, 307, 154, 107
+};
 
 /*
  * Client data (each client gets its own)
@@ -164,6 +176,8 @@ struct lm63_data {
 			   1: local high limit
 			   2: remote critical limit
 			   3-14: lookup table */
+	s16 temp1_input; /* local input for chips with extended resolution */
+	s16 temp1_max; /* local limit for chips with extended resolution */
 	s16 temp11[4];	/* 0: remote input
 			   1: remote low limit
 			   2: remote high limit
@@ -229,6 +243,7 @@ static struct lm63_data *lm63_update_device(struct device *dev)
 	struct lm63_data *data = dev_get_drvdata(dev);
 	struct i2c_client *client = data->client;
 	unsigned long next_update;
+	int msb, lsb;
 
 	mutex_lock(&data->update_lock);
 
@@ -253,10 +268,26 @@ static struct lm63_data *lm63_update_device(struct device *dev)
 		data->pwm1[0] = i2c_smbus_read_byte_data(client,
 				LM63_REG_PWM_VALUE);
 
-		data->temp8[0] = i2c_smbus_read_byte_data(client,
-				 LM63_REG_LOCAL_TEMP);
-		data->temp8[1] = i2c_smbus_read_byte_data(client,
-				 LM63_REG_LOCAL_HIGH);
+		if (data->kind == ctf2301) {
+			msb = i2c_smbus_read_byte_data(client,
+						       LM63_REG_LOCAL_TEMP);
+			lsb = i2c_smbus_read_byte_data(client,
+						       CTF2301_REG_LOCAL_TEMP_LSB);
+			if (msb >= 0 && lsb >= 0)
+				data->temp1_input = ((u16)(u8)msb << 8) | lsb;
+
+			msb = i2c_smbus_read_byte_data(client,
+						       LM63_REG_LOCAL_HIGH);
+			lsb = i2c_smbus_read_byte_data(client,
+						       CTF2301_REG_LOCAL_HIGH_LSB);
+			if (msb >= 0 && lsb >= 0)
+				data->temp1_max = ((u16)(u8)msb << 8) | lsb;
+		} else {
+			data->temp8[0] = i2c_smbus_read_byte_data(client,
+								  LM63_REG_LOCAL_TEMP);
+			data->temp8[1] = i2c_smbus_read_byte_data(client,
+								  LM63_REG_LOCAL_HIGH);
+		}
 
 		/* order matters for temp2_input */
 		data->temp11[0] = i2c_smbus_read_byte_data(client,
@@ -276,7 +307,7 @@ static struct lm63_data *lm63_update_device(struct device *dev)
 				| i2c_smbus_read_byte_data(client,
 				  LM63_REG_REMOTE_OFFSET_LSB);
 
-		if (data->kind == lm96163)
+		if (data->kind == lm96163 || data->kind == ctf2301)
 			data->temp11u = (i2c_smbus_read_byte_data(client,
 					LM96163_REG_REMOTE_TEMP_U_MSB) << 8)
 				      | i2c_smbus_read_byte_data(client,
@@ -366,6 +397,49 @@ static ssize_t set_fan(struct device *dev, struct device_attribute *dummy,
 	return count;
 }
 
+static int lm63_pwm_from_reg(struct lm63_data *data, u8 reg)
+{
+	if (data->pwm_highres)
+		return reg;
+
+	return reg >= 2 * data->pwm1_freq ?
+	       LM63_MAX_PWM :
+	       (reg * LM63_MAX_PWM + data->pwm1_freq) / (2 * data->pwm1_freq);
+}
+
+static u8 lm63_pwm_to_reg(struct lm63_data *data, unsigned long val)
+{
+	val = clamp_val(val, 0, LM63_MAX_PWM);
+
+	if (data->pwm_highres)
+		return val;
+
+	return (val * data->pwm1_freq * 2 + LM63_MAX_PWM / 2) / LM63_MAX_PWM;
+}
+
+static void lm63_update_pwm_highres(struct lm63_data *data,
+				    int config_enhanced)
+{
+	data->pwm_highres = !(data->config_fan & 0x08) &&
+			     data->pwm1_freq == 8 &&
+			     (config_enhanced & 0x10);
+}
+
+static int lm63_write_pwm(struct lm63_data *data, int nr, unsigned long val)
+{
+	struct i2c_client *client = data->client;
+	u8 reg = nr ? LM63_REG_LUT_PWM(nr - 1) : LM63_REG_PWM_VALUE;
+	u8 regval = lm63_pwm_to_reg(data, val);
+	int ret;
+
+	ret = i2c_smbus_write_byte_data(client, reg, regval);
+	if (ret)
+		return ret;
+	data->pwm1[nr] = regval;
+
+	return 0;
+}
+
 static ssize_t show_pwm1(struct device *dev, struct device_attribute *devattr,
 			 char *buf)
 {
@@ -375,12 +449,7 @@ static ssize_t show_pwm1(struct device *dev, struct device_attribute *devattr,
 	int pwm;
 
 	mutex_lock(&data->update_lock);
-	if (data->pwm_highres)
-		pwm = data->pwm1[nr];
-	else
-		pwm = data->pwm1[nr] >= 2 * data->pwm1_freq ?
-		       255 : (data->pwm1[nr] * 255 + data->pwm1_freq) /
-		       (2 * data->pwm1_freq);
+	pwm = lm63_pwm_from_reg(data, data->pwm1[nr]);
 	mutex_unlock(&data->update_lock);
 
 	return sprintf(buf, "%d\n", pwm);
@@ -391,11 +460,9 @@ static ssize_t set_pwm1(struct device *dev, struct device_attribute *devattr,
 {
 	struct sensor_device_attribute *attr = to_sensor_dev_attr(devattr);
 	struct lm63_data *data = dev_get_drvdata(dev);
-	struct i2c_client *client = data->client;
 	int nr = attr->index;
 	unsigned long val;
 	int err;
-	u8 reg;
 
 	if (!(data->config_fan & 0x20)) /* register is read-only */
 		return -EPERM;
@@ -404,15 +471,11 @@ static ssize_t set_pwm1(struct device *dev, struct device_attribute *devattr,
 	if (err)
 		return err;
 
-	reg = nr ? LM63_REG_LUT_PWM(nr - 1) : LM63_REG_PWM_VALUE;
-	val = clamp_val(val, 0, 255);
-
 	mutex_lock(&data->update_lock);
-	data->pwm1[nr] = data->pwm_highres ? val :
-			(val * data->pwm1_freq * 2 + 127) / 255;
-	i2c_smbus_write_byte_data(client, reg, data->pwm1[nr]);
+	err = lm63_write_pwm(data, nr, val);
 	mutex_unlock(&data->update_lock);
-	return count;
+
+	return err ? err : count;
 }
 
 static ssize_t pwm1_enable_show(struct device *dev,
@@ -485,7 +548,9 @@ static ssize_t pwm1_freq_store(struct device *dev,
 	struct lm63_data *data = dev_get_drvdata(dev);
 	struct i2c_client *client = data->client;
 	unsigned long val, pfr_fast, pfr_slow, err_fast, err_slow, pfr;
+	u8 config_fan;
 	bool slow_clock;
+	int config_enhanced = 0;
 	int ret;
 
 	ret = kstrtoul(buf, 10, &val);
@@ -511,35 +576,58 @@ static ssize_t pwm1_freq_store(struct device *dev,
 
 	mutex_lock(&data->update_lock);
 	ret = i2c_smbus_read_byte_data(client, LM63_REG_CONFIG_FAN);
-	if (ret < 0) {
-		mutex_unlock(&data->update_lock);
-		return ret;
-	}
-	data->config_fan = ret;
+	if (ret < 0)
+		goto unlock;
+	config_fan = ret;
+	data->config_fan = config_fan;
 
-	if (!(data->config_fan & 0x20)) { /* register is read-only */
-		mutex_unlock(&data->update_lock);
-		return -EPERM;
+	if (!(config_fan & 0x20)) { /* register is read-only */
+		ret = -EPERM;
+		goto unlock;
 	}
 
-	if (data->kind == lm96163) {
-		ret = i2c_smbus_read_byte_data(client, LM96163_REG_CONFIG_ENHANCED);
-		if (ret < 0) {
-			mutex_unlock(&data->update_lock);
-			return ret;
+	if (data->kind == lm96163 || data->kind == ctf2301) {
+		ret = i2c_smbus_read_byte_data(client,
+					       LM96163_REG_CONFIG_ENHANCED);
+		if (ret < 0)
+			goto unlock;
+		config_enhanced = ret;
+		lm63_update_pwm_highres(data, config_enhanced);
+
+		if (data->kind == ctf2301) {
+			if (!slow_clock && pfr == 8)
+				config_enhanced |= 0x10;
+			else
+				config_enhanced &= ~0x10;
+			ret = i2c_smbus_write_byte_data(client,
+							LM96163_REG_CONFIG_ENHANCED,
+							config_enhanced);
+			if (ret)
+				goto unlock;
+			lm63_update_pwm_highres(data, config_enhanced);
 		}
-		data->pwm_highres = !slow_clock && pfr == 8 && (ret & 0x10);
 	}
 
 	if (slow_clock)
-		data->config_fan |= 0x08;
+		config_fan |= 0x08;
 	else
-		data->config_fan &= ~0x08;
-	i2c_smbus_write_byte_data(client, LM63_REG_CONFIG_FAN, data->config_fan);
-	i2c_smbus_write_byte_data(client, LM63_REG_PWM_FREQ, pfr);
+		config_fan &= ~0x08;
+	ret = i2c_smbus_write_byte_data(client, LM63_REG_PWM_FREQ, pfr);
+	if (ret)
+		goto unlock;
 	data->pwm1_freq = pfr;
+	lm63_update_pwm_highres(data, config_enhanced);
+	ret = i2c_smbus_write_byte_data(client, LM63_REG_CONFIG_FAN,
+					config_fan);
+	if (ret)
+		goto unlock;
+
+	data->config_fan = config_fan;
+	lm63_update_pwm_highres(data, config_enhanced);
+unlock:
 	mutex_unlock(&data->update_lock);
-	return count;
+
+	return ret ? ret : count;
 }
 
 /*
@@ -554,7 +642,17 @@ static ssize_t show_local_temp8(struct device *dev,
 {
 	struct sensor_device_attribute *attr = to_sensor_dev_attr(devattr);
 	struct lm63_data *data = lm63_update_device(dev);
-	return sprintf(buf, "%d\n", TEMP8_FROM_REG(data->temp8[attr->index]));
+	int temp;
+
+	mutex_lock(&data->update_lock);
+	if (data->kind == ctf2301)
+		temp = TEMP12_FROM_REG(attr->index ? data->temp1_max :
+				       data->temp1_input);
+	else
+		temp = TEMP8_FROM_REG(data->temp8[attr->index]);
+	mutex_unlock(&data->update_lock);
+
+	return sprintf(buf, "%d\n", temp);
 }
 
 static ssize_t show_remote_temp8(struct device *dev,
@@ -594,6 +692,21 @@ static ssize_t set_temp8(struct device *dev, struct device_attribute *devattr,
 		return err;
 
 	mutex_lock(&data->update_lock);
+	if (data->kind == ctf2301 && nr == 1) {
+		temp = TEMP12_TO_REG(val);
+		err = i2c_smbus_write_byte_data(client, LM63_REG_LOCAL_HIGH,
+						temp >> 8);
+		if (!err)
+			err = i2c_smbus_write_byte_data(client,
+							CTF2301_REG_LOCAL_HIGH_LSB,
+							temp & 0xf0);
+		if (!err)
+			data->temp1_max = temp;
+		mutex_unlock(&data->update_lock);
+
+		return err ? err : count;
+	}
+
 	switch (nr) {
 	case 2:
 		reg = LM63_REG_REMOTE_TCRIT;
@@ -610,10 +723,12 @@ static ssize_t set_temp8(struct device *dev, struct device_attribute *devattr,
 		reg = LM63_REG_LUT_TEMP(nr - 3);
 		temp = lut_temp_to_reg(data, val);
 	}
-	data->temp8[nr] = temp;
-	i2c_smbus_write_byte_data(client, reg, temp);
+	err = i2c_smbus_write_byte_data(client, reg, temp);
+	if (!err)
+		data->temp8[nr] = temp;
 	mutex_unlock(&data->update_lock);
-	return count;
+
+	return err ? err : count;
 }
 
 static ssize_t show_temp11(struct device *dev, struct device_attribute *devattr,
@@ -787,11 +902,40 @@ static ssize_t temp2_crit_hyst_store(struct device *dev,
  * Set conversion rate.
  * client->update_lock must be held when calling this function.
  */
-static void lm63_set_convrate(struct lm63_data *data, unsigned int interval)
+static unsigned int lm63_update_interval(struct lm63_data *data, u8 rate)
+{
+	if (data->kind == ctf2301)
+		return ctf2301_update_intervals[min_t(unsigned int, rate,
+					ARRAY_SIZE(ctf2301_update_intervals) - 1)];
+
+	return UPDATE_INTERVAL(data->max_convrate_hz, rate);
+}
+
+static int lm63_set_convrate(struct lm63_data *data, unsigned int interval)
 {
 	struct i2c_client *client = data->client;
-	unsigned int update_interval;
-	int i;
+	unsigned int update_interval, error, best_error = ~0U;
+	unsigned int best = 0;
+	unsigned int i;
+	int ret;
+
+	if (data->kind == ctf2301) {
+		for (i = 0; i < ARRAY_SIZE(ctf2301_update_intervals); i++) {
+			error = abs_diff(interval, ctf2301_update_intervals[i]);
+			if (error < best_error) {
+				best_error = error;
+				best = i;
+			}
+		}
+
+		ret = i2c_smbus_write_byte_data(client, LM63_REG_CONVRATE,
+						best);
+		if (ret)
+			return ret;
+		data->update_interval = ctf2301_update_intervals[best];
+
+		return 0;
+	}
 
 	/* Shift calculations to avoid rounding errors */
 	interval <<= 6;
@@ -803,8 +947,12 @@ static void lm63_set_convrate(struct lm63_data *data, unsigned int interval)
 		if (interval >= update_interval * 3 / 4)
 			break;
 
-	i2c_smbus_write_byte_data(client, LM63_REG_CONVRATE, i);
+	ret = i2c_smbus_write_byte_data(client, LM63_REG_CONVRATE, i);
+	if (ret)
+		return ret;
 	data->update_interval = UPDATE_INTERVAL(data->max_convrate_hz, i);
+
+	return 0;
 }
 
 static ssize_t update_interval_show(struct device *dev,
@@ -828,10 +976,10 @@ static ssize_t update_interval_store(struct device *dev,
 		return err;
 
 	mutex_lock(&data->update_lock);
-	lm63_set_convrate(data, clamp_val(val, 0, 100000));
+	err = lm63_set_convrate(data, clamp_val(val, 0, 100000));
 	mutex_unlock(&data->update_lock);
 
-	return count;
+	return err ? err : count;
 }
 
 static ssize_t temp2_type_show(struct device *dev,
@@ -996,6 +1144,75 @@ static DEVICE_ATTR_RO(alarms);
 
 static DEVICE_ATTR_RW(update_interval);
 
+static int lm63_get_max_state(struct thermal_cooling_device *cdev,
+			      unsigned long *state)
+{
+	*state = LM63_MAX_PWM;
+
+	return 0;
+}
+
+static int lm63_get_cur_state(struct thermal_cooling_device *cdev,
+			      unsigned long *state)
+{
+	struct lm63_data *data = cdev->devdata;
+	int ret;
+
+	mutex_lock(&data->update_lock);
+	ret = i2c_smbus_read_byte_data(data->client, LM63_REG_PWM_VALUE);
+	if (ret < 0)
+		goto unlock;
+
+	data->pwm1[0] = ret;
+	*state = lm63_pwm_from_reg(data, data->pwm1[0]);
+unlock:
+	mutex_unlock(&data->update_lock);
+
+	return ret < 0 ? ret : 0;
+}
+
+static int lm63_set_cur_state(struct thermal_cooling_device *cdev,
+			      unsigned long state)
+{
+	struct lm63_data *data = cdev->devdata;
+	int config_fan;
+	int ret;
+
+	if (state > LM63_MAX_PWM)
+		return -EINVAL;
+
+	mutex_lock(&data->update_lock);
+	config_fan = i2c_smbus_read_byte_data(data->client,
+					      LM63_REG_CONFIG_FAN);
+	if (config_fan < 0) {
+		ret = config_fan;
+		goto unlock;
+	}
+	data->config_fan = config_fan;
+
+	if (!(config_fan & 0x20)) {
+		config_fan |= 0x20;
+		ret = i2c_smbus_write_byte_data(data->client,
+						LM63_REG_CONFIG_FAN,
+						config_fan);
+		if (ret)
+			goto unlock;
+		data->config_fan = config_fan;
+	}
+
+	ret = lm63_write_pwm(data, 0, state);
+unlock:
+	mutex_unlock(&data->update_lock);
+
+	return ret;
+}
+
+static const struct thermal_cooling_device_ops lm63_cooling_ops = {
+	.get_max_state = lm63_get_max_state,
+	.get_cur_state = lm63_get_cur_state,
+	.set_cur_state = lm63_set_cur_state,
+};
+
 static struct attribute *lm63_attributes[] = {
 	&sensor_dev_attr_pwm1.dev_attr.attr,
 	&dev_attr_pwm1_enable.attr,
@@ -1078,8 +1295,8 @@ static const struct attribute_group lm63_group_extra_lut = {
  * On LM63, temp2_crit can be set only once, which should be job
  * of the bootloader.
  * On LM64, temp2_crit can always be set.
- * On LM96163, temp2_crit can be set if bit 1 of the configuration
- * register is true.
+ * On LM96163 and CTF2301, temp2_crit can be set if bit 1 of the
+ * configuration register is true.
  */
 static umode_t lm63_attribute_mode(struct kobject *kobj,
 				   struct attribute *attr, int index)
@@ -1089,7 +1306,8 @@ static umode_t lm63_attribute_mode(struct kobject *kobj,
 
 	if (attr == &sensor_dev_attr_temp2_crit.dev_attr.attr
 	    && (data->kind == lm64 ||
-		(data->kind == lm96163 && (data->config & 0x02))))
+		((data->kind == lm96163 || data->kind == ctf2301) &&
+		 (data->config & 0x02))))
 		return attr->mode | S_IWUSR;
 
 	return attr->mode;
@@ -1164,29 +1382,48 @@ static int lm63_detect(struct i2c_client *client,
  * Ideally we shouldn't have to initialize anything, since the BIOS
  * should have taken care of everything
  */
-static void lm63_init_client(struct lm63_data *data)
+static int lm63_init_client(struct lm63_data *data)
 {
 	struct i2c_client *client = data->client;
 	struct device *dev = &client->dev;
 	u8 convrate;
+	int ret;
 
-	data->config = i2c_smbus_read_byte_data(client, LM63_REG_CONFIG1);
-	data->config_fan = i2c_smbus_read_byte_data(client,
-						    LM63_REG_CONFIG_FAN);
+	ret = i2c_smbus_read_byte_data(client, LM63_REG_CONFIG1);
+	if (ret < 0)
+		return ret;
+	data->config = ret;
+	ret = i2c_smbus_read_byte_data(client, LM63_REG_CONFIG_FAN);
+	if (ret < 0)
+		return ret;
+	data->config_fan = ret;
 
 	/* Start converting if needed */
 	if (data->config & 0x40) { /* standby */
 		dev_dbg(dev, "Switching to operational mode\n");
 		data->config &= 0xA7;
-		i2c_smbus_write_byte_data(client, LM63_REG_CONFIG1,
-					  data->config);
+		ret = i2c_smbus_write_byte_data(client, LM63_REG_CONFIG1,
+						data->config);
+		if (ret)
+			return ret;
 	}
 	/* Tachometer is always enabled on LM64 */
 	if (data->kind == lm64)
 		data->config |= 0x04;
+	if (data->kind == ctf2301 && !client->irq && !(data->config & 0x04)) {
+		dev_dbg(dev, "Configuring ALERT/TACH pin for tachometer input\n");
+		data->config |= 0x04;
+		ret = i2c_smbus_write_byte_data(client, LM63_REG_CONFIG1,
+						data->config);
+		if (ret)
+			return ret;
+	}
 
 	/* We may need pwm1_freq before ever updating the client data */
-	data->pwm1_freq = i2c_smbus_read_byte_data(client, LM63_REG_PWM_FREQ) & 0x1f;
+	ret = i2c_smbus_read_byte_data(client, LM63_REG_PWM_FREQ);
+	if (ret < 0)
+		return ret;
+	data->pwm1_freq = ret & 0x1f;
 	if (data->pwm1_freq == 0)
 		data->pwm1_freq = 1;
 
@@ -1199,25 +1436,38 @@ static void lm63_init_client(struct lm63_data *data)
 	case lm96163:
 		data->max_convrate_hz = LM96163_MAX_CONVRATE_HZ;
 		data->lut_size = 12;
-		data->trutherm
-		  = i2c_smbus_read_byte_data(client,
-					     LM96163_REG_TRUTHERM) & 0x02;
+		ret = i2c_smbus_read_byte_data(client, LM96163_REG_TRUTHERM);
+		if (ret < 0)
+			return ret;
+		data->trutherm = ret & 0x02;
+		break;
+	case ctf2301:
+		data->lut_size = 12;
 		break;
 	}
-	convrate = i2c_smbus_read_byte_data(client, LM63_REG_CONVRATE);
-	if (unlikely(convrate > LM63_MAX_CONVRATE))
+	ret = i2c_smbus_read_byte_data(client, LM63_REG_CONVRATE);
+	if (ret < 0)
+		return ret;
+	convrate = ret;
+	if (data->kind == ctf2301 &&
+	    convrate >= ARRAY_SIZE(ctf2301_update_intervals))
+		convrate = ARRAY_SIZE(ctf2301_update_intervals) - 1;
+	else if (unlikely(convrate > LM63_MAX_CONVRATE))
 		convrate = LM63_MAX_CONVRATE;
-	data->update_interval = UPDATE_INTERVAL(data->max_convrate_hz,
-						convrate);
+	data->update_interval = lm63_update_interval(data, convrate);
 
 	/*
-	 * For LM96163, check if high resolution PWM
+	 * For LM96163 and CTF2301, check if high resolution PWM
 	 * and unsigned temperature format is enabled.
 	 */
-	if (data->kind == lm96163) {
-		u8 config_enhanced
-		  = i2c_smbus_read_byte_data(client,
-					     LM96163_REG_CONFIG_ENHANCED);
+	if (data->kind == lm96163 || data->kind == ctf2301) {
+		u8 config_enhanced;
+
+		ret = i2c_smbus_read_byte_data(client,
+					       LM96163_REG_CONFIG_ENHANCED);
+		if (ret < 0)
+			return ret;
+		config_enhanced = ret;
 		if (config_enhanced & 0x20)
 			data->lut_temp_highres = true;
 		if ((config_enhanced & 0x10)
@@ -1238,6 +1488,8 @@ static void lm63_init_client(struct lm63_data *data)
 	dev_dbg(dev, "PWM output active %s, %s mode\n",
 		(data->config_fan & 0x10) ? "low" : "high",
 		(data->config_fan & 0x20) ? "manual" : "auto");
+
+	return 0;
 }
 
 static const struct i2c_device_id lm63_id[];
@@ -1245,9 +1497,11 @@ static const struct i2c_device_id lm63_id[];
 static int lm63_probe(struct i2c_client *client)
 {
 	struct device *dev = &client->dev;
+	struct thermal_cooling_device *cdev;
 	struct device *hwmon_dev;
 	struct lm63_data *data;
 	int groups = 0;
+	int ret;
 
 	data = devm_kzalloc(dev, sizeof(struct lm63_data), GFP_KERNEL);
 	if (!data)
@@ -1262,7 +1516,9 @@ static int lm63_probe(struct i2c_client *client)
 		data->temp2_offset = 16000;
 
 	/* Initialize chip */
-	lm63_init_client(data);
+	ret = lm63_init_client(data);
+	if (ret)
+		return dev_err_probe(dev, ret, "failed to initialize device\n");
 
 	/* Register sysfs hooks */
 	data->groups[groups++] = &lm63_group;
@@ -1272,11 +1528,26 @@ static int lm63_probe(struct i2c_client *client)
 	if (data->kind == lm96163) {
 		data->groups[groups++] = &lm63_group_temp2_type;
 		data->groups[groups++] = &lm63_group_extra_lut;
+	} else if (data->kind == ctf2301) {
+		data->groups[groups++] = &lm63_group_extra_lut;
 	}
 
 	hwmon_dev = devm_hwmon_device_register_with_groups(dev, client->name,
 							   data, data->groups);
-	return PTR_ERR_OR_ZERO(hwmon_dev);
+	if (IS_ERR(hwmon_dev))
+		return PTR_ERR(hwmon_dev);
+
+	if (IS_ENABLED(CONFIG_THERMAL_OF) &&
+	    of_property_present(dev->of_node, "#cooling-cells")) {
+		cdev = devm_thermal_of_cooling_device_register(dev, 0,
+							       client->name,
+							       data,
+							       &lm63_cooling_ops);
+		if (IS_ERR(cdev))
+			return PTR_ERR(cdev);
+	}
+
+	return 0;
 }
 
 /*
@@ -1287,6 +1558,7 @@ static const struct i2c_device_id lm63_id[] = {
 	{ .name = "lm63", .driver_data = lm63 },
 	{ .name = "lm64", .driver_data = lm64 },
 	{ .name = "lm96163", .driver_data = lm96163 },
+	{ .name = "ctf2301", .driver_data = ctf2301 },
 	{ }
 };
 MODULE_DEVICE_TABLE(i2c, lm63_id);
@@ -1304,6 +1576,10 @@ static const struct of_device_id __maybe_unused lm63_of_match[] = {
 		.compatible = "national,lm96163",
 		.data = (void *)lm96163
 	},
+	{
+		.compatible = "sensylink,ctf2301",
+		.data = (void *)ctf2301
+	},
 	{ },
 };
 MODULE_DEVICE_TABLE(of, lm63_of_match);

-- 
2.55.0

Re: [PATCH v4 4/6] hwmon: (lm63) Add Sensylink CTF2301 support

[sashiko-bot]

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [Medium] TEMP12_TO_REG and software caching logic fail to align with 12-bit hardware truncation, causing sysfs cache inconsistency.
--

commit c4469f3bd5935eebd0d31fe1d782461f5b52c354
Author: Troy Mitchell <troy.mitchell@linux.dev>

hwmon: (lm63) Add Sensylink CTF2301 support

This commit adds support for the Sensylink CTF2301 to the LM63 driver. It
registers an optional thermal cooling device when the device tree node
provides #cooling-cells and implements the specific temperature scaling for
the CTF2301 chip.

> diff --git a/drivers/hwmon/lm63.c b/drivers/hwmon/lm63.c
> --- a/drivers/hwmon/lm63.c
> +++ b/drivers/hwmon/lm63.c
[ ... ]
> @@ -121,6 +126,9 @@ static const unsigned short normal_i2c[] = { 0x18, 0x4c, 0x4e, I2C_CLIENT_END };
>  							    127000), 1000)
>  #define TEMP8U_TO_REG(val)	DIV_ROUND_CLOSEST(clamp_val((val), 0, \
>  							    255000), 1000)
> +#define TEMP12_FROM_REG(reg)	DIV_ROUND_CLOSEST((s16)(reg) * 1000, 256)
> +#define TEMP12_TO_REG(val)	DIV_ROUND_CLOSEST(clamp_val((val), -128000, \
> +							     127937) * 256, 1000)
>  #define TEMP11_FROM_REG(reg)	((reg) / 32 * 125)
>  #define TEMP11_TO_REG(val)	(DIV_ROUND_CLOSEST(clamp_val((val), -128000, \
>  							     127875), 125) * 32)
[ ... ]
> @@ -594,6 +692,21 @@ static ssize_t set_temp8(struct device *dev, struct device_attribute *devattr,
>  		return err;
>  
>  	mutex_lock(&data->update_lock);
> +	if (data->kind == ctf2301 && nr == 1) {
> +		temp = TEMP12_TO_REG(val);
> +		err = i2c_smbus_write_byte_data(client, LM63_REG_LOCAL_HIGH,
> +						temp >> 8);
> +		if (!err)
> +			err = i2c_smbus_write_byte_data(client,
> +							CTF2301_REG_LOCAL_HIGH_LSB,
> +							temp & 0xf0);
> +		if (!err)
> +			data->temp1_max = temp;

[Severity: Medium]
Does this sequence cause the sysfs cache to become inconsistent with the
hardware state?

The TEMP12_TO_REG macro scales the value but does not align or quantize it
to a 12-bit hardware boundary, leaving fractional bits present in the lowest
4 bits. Here in set_temp8, those lowest 4 bits of temp are masked out with
0xf0 when writing to the hardware register, but the unmasked temp is saved in
data->temp1_max.

If a user writes a temperature limit that is not a perfect multiple of 0.0625
degrees Celsius, will this cause the sysfs interface to return the
un-truncated value immediately after the write, but later shift to the
truncated value upon the next device poll?

> +		mutex_unlock(&data->update_lock);
> +
> +		return err ? err : count;
> +	}
> +

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260831-ctl2301-v4-0-b335d52a0570@linux.dev?part=4

Re: [PATCH v4 4/6] hwmon: (lm63) Add Sensylink CTF2301 support

[Troy Mitchell]

[-- Attachment #1: Type: text/plain, Size: 2670 bytes --]

On Mon, Aug 31, 2026 at 04:32:56PM +0000, sashiko-bot@kernel.org wrote:
> Thank you for your contribution! Sashiko AI review found 1 potential
> issue(s) to consider:
> - [Medium] TEMP12_TO_REG and software caching logic fail to align with
> 12-bit hardware truncation, causing sysfs cache inconsistency.
>
> [...]
>
> @@ -121,6 +126,9 @@ static const unsigned short normal_i2c[] = { 0x18, 0x4c, 0x4e, I2C_CLIENT_END };
>  							    127000), 1000)
>  #define TEMP8U_TO_REG(val)	DIV_ROUND_CLOSEST(clamp_val((val), 0, \
>  							    255000), 1000)
> +#define TEMP12_FROM_REG(reg)	DIV_ROUND_CLOSEST((s16)(reg) * 1000, 256)
> +#define TEMP12_TO_REG(val)	DIV_ROUND_CLOSEST(clamp_val((val), -128000, \
> +							     127937) * 256, 1000)
>  #define TEMP11_FROM_REG(reg)	((reg) / 32 * 125)
>  #define TEMP11_TO_REG(val)	(DIV_ROUND_CLOSEST(clamp_val((val), -128000, \
>  							     127875), 125) * 32)
>
> [...]
>
> @@ -594,6 +692,21 @@ static ssize_t set_temp8(struct device *dev, struct device_attribute *devattr,
>  		return err;
>  
>  	mutex_lock(&data->update_lock);
> +	if (data->kind == ctf2301 && nr == 1) {
> +		temp = TEMP12_TO_REG(val);
> +		err = i2c_smbus_write_byte_data(client, LM63_REG_LOCAL_HIGH,
> +						temp >> 8);
> +		if (!err)
> +			err = i2c_smbus_write_byte_data(client,
> +							CTF2301_REG_LOCAL_HIGH_LSB,
> +							temp & 0xf0);
> +		if (!err)
> +			data->temp1_max = temp;
>
> [Severity: Medium]
> Does this sequence cause the sysfs cache to become inconsistent with the
> hardware state?
>
> The TEMP12_TO_REG macro scales the value but does not align or quantize it
> to a 12-bit hardware boundary, leaving fractional bits present in the
> lowest 4 bits. Here in set_temp8, those lowest 4 bits of temp are masked
> out with 0xf0 when writing to the hardware register, but the unmasked temp
> is saved in data->temp1_max.
>
> If a user writes a temperature limit that is not a perfect multiple of
> 0.0625 degrees Celsius, will this cause the sysfs interface to return the
> un-truncated value immediately after the write, but later shift to the
> truncated value upon the next device poll?

This is a bug. TEMP12_TO_REG() currently produces an unaligned 8.8
fixed-point value. The register write discards bits 3:0, while the cache
retains them, so an immediate sysfs read can differ from the value read
back from the device after the next update.

I will fix this in the next version by quantizing the converted value to
the CTF2301's 0.0625 degree C resolution before using it for both the
register writes and the cache update.

                                            - Troy

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 273 bytes --]

[PATCH v4 5/6] riscv: dts: spacemit: Add K3 I2C6 pinctrl state

[Troy Mitchell]

Add the K3 I2C6 pinctrl state used by boards that route I2C6 to pads 7
and 8. Use a 25 mA drive strength, which is supported for both K3 IO
voltage domains, and leave board-specific IO voltage selection to board
files.

Signed-off-by: Troy Mitchell <troy.mitchell@linux.dev>
---
 arch/riscv/boot/dts/spacemit/k3-pinctrl.dtsi | 11 +++++++++++
 1 file changed, 11 insertions(+)

diff --git a/arch/riscv/boot/dts/spacemit/k3-pinctrl.dtsi b/arch/riscv/boot/dts/spacemit/k3-pinctrl.dtsi
index 6482d24be139..cb4f4f4e33af 100644
--- a/arch/riscv/boot/dts/spacemit/k3-pinctrl.dtsi
+++ b/arch/riscv/boot/dts/spacemit/k3-pinctrl.dtsi
@@ -101,6 +101,17 @@ i2c2-1-pins {
 		};
 	};
 
+	/omit-if-no-ref/
+	i2c6_1_cfg: i2c6-1-cfg {
+		i2c6-1-pins {
+			pinmux = <K3_PADCONF(7, 6)>,	/* i2c6 scl */
+				 <K3_PADCONF(8, 6)>;	/* i2c6 sda */
+
+			bias-pull-up = <0>;
+			drive-strength = <25>;
+		};
+	};
+
 	/omit-if-no-ref/
 	i2c8_cfg: i2c8-cfg {
 		i2c8-pins {

-- 
2.55.0

Re: [PATCH v4 5/6] riscv: dts: spacemit: Add K3 I2C6 pinctrl state

[sashiko-bot]

> Add the K3 I2C6 pinctrl state used by boards that route I2C6 to pads 7
> and 8. Use a 25 mA drive strength, which is supported for both K3 IO
> voltage domains, and leave board-specific IO voltage selection to board
> files.
> 
> Signed-off-by: Troy Mitchell <troy.mitchell@linux.dev>

Sashiko has reviewed this patch and found no issues. It looks great!

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260831-ctl2301-v4-0-b335d52a0570@linux.dev?part=5

[PATCH v4 6/6] riscv: dts: spacemit: Add CTF2301 on K3 CoM260

[Troy Mitchell]

Enable I2C6 on the K3 CoM260 module and describe the Sensylink
CTF2301 temperature sensor and fan controller at address 0x4c.

Override the I2C6 pin power source at board level because the IO rail
voltage is board-specific.

Signed-off-by: Troy Mitchell <troy.mitchell@linux.dev>
---
 arch/riscv/boot/dts/spacemit/k3-com260.dtsi | 17 +++++++++++++++++
 1 file changed, 17 insertions(+)

diff --git a/arch/riscv/boot/dts/spacemit/k3-com260.dtsi b/arch/riscv/boot/dts/spacemit/k3-com260.dtsi
index 37627963a649..92dea8639779 100644
--- a/arch/riscv/boot/dts/spacemit/k3-com260.dtsi
+++ b/arch/riscv/boot/dts/spacemit/k3-com260.dtsi
@@ -58,6 +58,23 @@ serial-number {
 	};
 };
 
+&i2c6_1_cfg {
+	i2c6-1-pins {
+		power-source = <3300>;
+	};
+};
+
+&i2c6 {
+	pinctrl-names = "default";
+	pinctrl-0 = <&i2c6_1_cfg>;
+	status = "okay";
+
+	temperature-sensor@4c {
+		compatible = "sensylink,ctf2301";
+		reg = <0x4c>;
+	};
+};
+
 &i2c8 {
 	pinctrl-names = "default";
 	pinctrl-0 = <&i2c8_cfg>;

-- 
2.55.0

Re: [PATCH v4 6/6] riscv: dts: spacemit: Add CTF2301 on K3 CoM260

[sashiko-bot]

> Enable I2C6 on the K3 CoM260 module and describe the Sensylink
> CTF2301 temperature sensor and fan controller at address 0x4c.
> 
> Override the I2C6 pin power source at board level because the IO rail
> voltage is board-specific.
> 
> Signed-off-by: Troy Mitchell <troy.mitchell@linux.dev>

Sashiko has reviewed this patch and found no issues. It looks great!

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260831-ctl2301-v4-0-b335d52a0570@linux.dev?part=6