[PATCH 0/5] hwmon: (pmbus) Add support for MAX20826 and similar devices

[PATCH 0/5] hwmon: (pmbus) Add support for MAX20826 and similar devices

[Nuno Sá via B4 Relay]

Add support for Analog Devices MAX20826, MAX20855B, MAX20908,
MAX20912 and MAX20916 dual-output multiphase PMBus voltage regulator
controllers.

This follows up on the earlier MAX20912/MAX20916 submission and the
review discussion around it [1]. Compared to that minimal driver, this
series covers the full MAX20826 family and implements the device-specific
pieces discussed there: optional EN GPIO handling, non-standard phase
reporting, direct address mode, and regulator support.

The first patches prepare/fix PMBus core support needed by the driver:
fix regulator enable/disable return handling, allow up to 16 phases, and
add a device-specific block-read hook. The remaining patches add the DT
binding and the MAX20826-family driver/documentation.

Guenter,

We are still checking the possibility of getting you docs for at least
one of the parts supported byu this series.

[1]: https://lore.kernel.org/linux-hwmon/20260707122701.751878-3-fredchen.openbmc@gmail.com/

---
Nuno Sá (5):
      hwmon: (pmbus/core) fix regulator enable/disable
      hwmon: (pmbus/core) increase number of phases and add new mask
      hwmon: (pmbus/core) Add mapping function to pmbus_read_block_data()
      dt-bindings: hwmon/pmbus: Document MAX20826 and similar devices
      hwmon: (pmbus) add support for MAX20826 and similar devices

 .../bindings/hwmon/pmbus/adi,max20826.yaml         |   56 ++
 Documentation/hwmon/index.rst                      |    1 +
 Documentation/hwmon/max20826.rst                   |  124 +++
 MAINTAINERS                                        |    9 +
 drivers/hwmon/pmbus/Kconfig                        |   19 +
 drivers/hwmon/pmbus/Makefile                       |    1 +
 drivers/hwmon/pmbus/max20826.c                     | 1037 ++++++++++++++++++++
 drivers/hwmon/pmbus/pmbus.h                        |    6 +-
 drivers/hwmon/pmbus/pmbus_core.c                   |   39 +-
 9 files changed, 1286 insertions(+), 6 deletions(-)
---
base-commit: 92413f439d1ec5e55b73ede8d66a7b971cbd1ced
change-id: 20260728-hwmon-max20826-support-c82ee65118ec
--

Thanks!
- Nuno Sá

[PATCH 1/5] hwmon: (pmbus/core) fix regulator enable/disable

[Nuno Sá via B4 Relay]

From: Nuno Sá <nuno.sa@analog.com>

pmbus_update_byte_data() can return the value of PMBUS_OPERATION which
can be different than 0 and that can mess with the regulator core
given _regulator_disable() explicitly checks for ret == 0 in order to
call _regulator_handle_consumer_disable().

Fixes: ddbb4db4ced1 ("hwmon: (pmbus) Add regulator support")
Signed-off-by: Nuno Sá <nuno.sa@analog.com>
---
 drivers/hwmon/pmbus/pmbus_core.c | 15 ++++++++++++---
 1 file changed, 12 insertions(+), 3 deletions(-)

diff --git a/drivers/hwmon/pmbus/pmbus_core.c b/drivers/hwmon/pmbus/pmbus_core.c
index 0081f16c3a95..7b58f7198574 100644
--- a/drivers/hwmon/pmbus/pmbus_core.c
+++ b/drivers/hwmon/pmbus/pmbus_core.c
@@ -3100,12 +3100,21 @@ static int _pmbus_regulator_on_off(struct regulator_dev *rdev, bool enable)
 	struct device *dev = rdev_get_dev(rdev);
 	struct i2c_client *client = to_i2c_client(dev->parent);
 	u8 page = rdev_get_id(rdev);
+	int rv;
 
 	guard(pmbus_lock)(client);
+	/*
+	 * pmbus_update_byte_data() can just return the value of
+	 * PMBUS_OPERATION and that's not what we want to return to the
+	 * regulator core.
+	 */
+	rv = pmbus_update_byte_data(client, page, PMBUS_OPERATION,
+				    PB_OPERATION_CONTROL_ON,
+				    enable ? PB_OPERATION_CONTROL_ON : 0);
+	if (rv < 0)
+		return rv;
 
-	return pmbus_update_byte_data(client, page, PMBUS_OPERATION,
-				      PB_OPERATION_CONTROL_ON,
-				      enable ? PB_OPERATION_CONTROL_ON : 0);
+	return 0;
 }
 
 static int pmbus_regulator_enable(struct regulator_dev *rdev)

-- 
2.55.0

Re: [PATCH 1/5] hwmon: (pmbus/core) fix regulator enable/disable

[sashiko-bot]

> From: Nuno Sá <nuno.sa@analog.com>
> 
> pmbus_update_byte_data() can return the value of PMBUS_OPERATION which
> can be different than 0 and that can mess with the regulator core
> given _regulator_disable() explicitly checks for ret == 0 in order to
> call _regulator_handle_consumer_disable().
> 
> Fixes: ddbb4db4ced1 ("hwmon: (pmbus) Add regulator support")
> Signed-off-by: Nuno Sá <nuno.sa@analog.com>

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

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260728-hwmon-max20826-support-v1-0-224766e0acd1@analog.com?part=1

Re: [PATCH 1/5] hwmon: (pmbus/core) fix regulator enable/disable

[Guenter Roeck]

On 7/28/26 09:03, Nuno Sá via B4 Relay wrote:
> From: Nuno Sá <nuno.sa@analog.com>
> 
> pmbus_update_byte_data() can return the value of PMBUS_OPERATION which
> can be different than 0 and that can mess with the regulator core
> given _regulator_disable() explicitly checks for ret == 0 in order to
> call _regulator_handle_consumer_disable().
> 
> Fixes: ddbb4db4ced1 ("hwmon: (pmbus) Add regulator support")
> Signed-off-by: Nuno Sá <nuno.sa@analog.com>

The real problem is that pmbus_update_byte_data() should only return 0
or an error code. The fix is in the "hwmon" branch and should be upstream
by the end of the week.

Guenter

[PATCH 2/5] hwmon: (pmbus/core) increase number of phases and add new mask

[Nuno Sá via B4 Relay]

From: Nuno Sá <nuno.sa@analog.com>

Increase the number of phases to 16 as a new upcoming device supports
such a number.

While at it, add a new mask for controlling the source of the output
voltage.

Signed-off-by: Nuno Sá <nuno.sa@analog.com>
---
 drivers/hwmon/pmbus/pmbus.h | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/drivers/hwmon/pmbus/pmbus.h b/drivers/hwmon/pmbus/pmbus.h
index 23e3eda58870..3d5586c67f84 100644
--- a/drivers/hwmon/pmbus/pmbus.h
+++ b/drivers/hwmon/pmbus/pmbus.h
@@ -242,6 +242,7 @@ enum pmbus_regs {
 /*
  * OPERATION
  */
+#define PB_OPERATION_CONTROL_V_SRC	GENMASK(5, 4)
 #define PB_OPERATION_CONTROL_ON		BIT(7)
 
 /*
@@ -386,7 +387,7 @@ enum pmbus_sensor_classes {
 };
 
 #define PMBUS_PAGES	32	/* Per PMBus specification */
-#define PMBUS_PHASES	10	/* Maximum number of phases per page */
+#define PMBUS_PHASES	16	/* Maximum number of phases per page */
 
 /* Functionality bit mask */
 #define PMBUS_HAVE_VIN		BIT(0)

-- 
2.55.0

Re: [PATCH 2/5] hwmon: (pmbus/core) increase number of phases and add new mask

[sashiko-bot]

> From: Nuno Sá <nuno.sa@analog.com>
> 
> Increase the number of phases to 16 as a new upcoming device supports
> such a number.
> 
> While at it, add a new mask for controlling the source of the output
> voltage.
> 
> Signed-off-by: Nuno Sá <nuno.sa@analog.com>

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

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260728-hwmon-max20826-support-v1-0-224766e0acd1@analog.com?part=2

[PATCH 3/5] hwmon: (pmbus/core) Add mapping function to pmbus_read_block_data()

[Nuno Sá via B4 Relay]

From: Nuno Sá <nuno.sa@analog.com>

This is in preparation for adding support to a device which needs to
use it's own read_block implementation.

Signed-off-by: Nuno Sá <nuno.sa@analog.com>
---
 drivers/hwmon/pmbus/pmbus.h      |  3 +++
 drivers/hwmon/pmbus/pmbus_core.c | 24 ++++++++++++++++++++++--
 2 files changed, 25 insertions(+), 2 deletions(-)

diff --git a/drivers/hwmon/pmbus/pmbus.h b/drivers/hwmon/pmbus/pmbus.h
index 3d5586c67f84..d697939ec892 100644
--- a/drivers/hwmon/pmbus/pmbus.h
+++ b/drivers/hwmon/pmbus/pmbus.h
@@ -459,6 +459,9 @@ struct pmbus_driver_info {
 	int (*read_byte_data)(struct i2c_client *client, int page, int reg);
 	int (*read_word_data)(struct i2c_client *client, int page, int phase,
 			      int reg);
+	/* size of data_buf is I2C_SMBUS_BLOCK_MAX + 2 */
+	int (*read_block_data)(struct i2c_client *client, int page, u8 reg,
+			       char *data_buf);
 	int (*write_byte_data)(struct i2c_client *client, int page, int reg,
 			      u8 byte);
 	int (*write_word_data)(struct i2c_client *client, int page, int reg,
diff --git a/drivers/hwmon/pmbus/pmbus_core.c b/drivers/hwmon/pmbus/pmbus_core.c
index 7b58f7198574..ff4572c473b7 100644
--- a/drivers/hwmon/pmbus/pmbus_core.c
+++ b/drivers/hwmon/pmbus/pmbus_core.c
@@ -533,6 +533,26 @@ static int pmbus_read_block_data(struct i2c_client *client, int page, u8 reg,
 	return rv;
 }
 
+/*
+ * _pmbus_read_block_data() is similar to pmbus_read_block_data(), but checks if
+ * a device specific mapping function exists and calls it if necessary.
+ */
+static int _pmbus_read_block_data(struct i2c_client *client, int page, u8 reg,
+				  char *data_buf)
+{
+	struct pmbus_data *data = i2c_get_clientdata(client);
+	const struct pmbus_driver_info *info = data->info;
+	int status;
+
+	if (info->read_block_data) {
+		status = info->read_block_data(client, page, reg, data_buf);
+		if (status != -ENODATA)
+			return status;
+	}
+
+	return pmbus_read_block_data(client, page, reg, data_buf);
+}
+
 static struct pmbus_sensor *pmbus_find_sensor(struct pmbus_data *data, int page,
 					      int reg)
 {
@@ -678,7 +698,7 @@ static bool __maybe_unused pmbus_check_block_register(struct i2c_client *client,
 	struct pmbus_data *data = i2c_get_clientdata(client);
 	char data_buf[I2C_SMBUS_BLOCK_MAX + 2];
 
-	rv = pmbus_read_block_data(client, page, reg, data_buf);
+	rv = _pmbus_read_block_data(client, page, reg, data_buf);
 	if (rv >= 0 && !(data->flags & PMBUS_SKIP_STATUS_CHECK))
 		rv = pmbus_check_status_cml(client);
 	if (rv < 0 && (data->flags & PMBUS_READ_STATUS_AFTER_FAILED_CHECK))
@@ -3564,7 +3584,7 @@ static ssize_t pmbus_debugfs_block_read(struct file *file, char __user *buf,
 	char data[I2C_SMBUS_BLOCK_MAX + 2] = { 0 };
 
 	scoped_guard(pmbus_lock, client) {
-		rc = pmbus_read_block_data(client, entry->page, entry->reg, data);
+		rc = _pmbus_read_block_data(client, entry->page, entry->reg, data);
 		if (rc < 0)
 			return rc;
 	}

-- 
2.55.0

Re: [PATCH 3/5] hwmon: (pmbus/core) Add mapping function to pmbus_read_block_data()

[sashiko-bot]

Thank you for your contribution! Sashiko AI review found 2 potential issue(s) to consider:

New issues:
- [High] Stack buffer overflow in `pmbus_debugfs_block_read` due to unbounded return value from custom `read_block_data` callback.

Pre-existing issues:
- [High] TOCTOU data race on PMBus hardware page selection during probe.
--

--- Patch [3]: [PATCH 3/5] hwmon: (pmbus/core) Add mapping function to pmbus_read_block_data() ---
commit c4bd4d03203a9fd14a7a2bcff69ec842b441fa68
Author: Nuno Sá <nuno.sa@analog.com>

hwmon: (pmbus/core) Add mapping function to pmbus_read_block_data()

This commit introduces a mapping function to pmbus_read_block_data() in
preparation for adding support to a device which needs to use its own
read_block implementation.

> diff --git a/drivers/hwmon/pmbus/pmbus.h b/drivers/hwmon/pmbus/pmbus.h
> index 3d5586c67f840..d697939ec8921 100644
> --- a/drivers/hwmon/pmbus/pmbus.h
> +++ b/drivers/hwmon/pmbus/pmbus.h
> @@ -459,6 +459,9 @@ struct pmbus_driver_info {
>  	int (*read_byte_data)(struct i2c_client *client, int page, int reg);
>  	int (*read_word_data)(struct i2c_client *client, int page, int phase,
>  			      int reg);
> +	/* size of data_buf is I2C_SMBUS_BLOCK_MAX + 2 */
> +	int (*read_block_data)(struct i2c_client *client, int page, u8 reg,
> +			       char *data_buf);

[Severity: High]
Does this API need a size limit parameter? The comment specifies the buffer
size is 34, without providing a limit parameter to the callback, which might
encourage callbacks to return values up to 34.

[ ... ]

> diff --git a/drivers/hwmon/pmbus/pmbus_core.c b/drivers/hwmon/pmbus/pmbus_core.c
> index 7b58f71985741..ff4572c473b70 100644
> --- a/drivers/hwmon/pmbus/pmbus_core.c
> +++ b/drivers/hwmon/pmbus/pmbus_core.c
[ ... ]
> @@ -678,7 +698,7 @@ static bool __maybe_unused pmbus_check_block_register(struct i2c_client *client,
>  	struct pmbus_data *data = i2c_get_clientdata(client);
>  	char data_buf[I2C_SMBUS_BLOCK_MAX + 2];
>  
> -	rv = pmbus_read_block_data(client, page, reg, data_buf);
> +	rv = _pmbus_read_block_data(client, page, reg, data_buf);

[Severity: High]
This is a pre-existing issue, but can this lead to a race condition with
sysfs accesses during probe?

Looking at pmbus_init_debugfs(), it traverses pages and reads registers
without taking the pmbus_lock. This indirectly invokes pmbus_set_page()
through pmbus_check_block_register():

pmbus_init_debugfs()
  pmbus_check_block_register()
    _pmbus_read_block_data()
      pmbus_read_block_data()
        pmbus_set_page()

If userspace reads or writes to a sysfs attribute concurrently, it acquires
the lock and sets the hardware page. Could the lockless call from
pmbus_init_debugfs() interleave between the sysfs thread's pmbus_set_page()
and its subsequent I2C operation, causing the sysfs thread to hit the wrong
PMBus page?

>  	if (rv >= 0 && !(data->flags & PMBUS_SKIP_STATUS_CHECK))
>  		rv = pmbus_check_status_cml(client);
[ ... ]
> @@ -3564,7 +3584,7 @@ static ssize_t pmbus_debugfs_block_read(struct file *file, char __user *buf,
>  	char data[I2C_SMBUS_BLOCK_MAX + 2] = { 0 };
>  
>  	scoped_guard(pmbus_lock, client) {
> -		rc = pmbus_read_block_data(client, entry->page, entry->reg, data);
> +		rc = _pmbus_read_block_data(client, entry->page, entry->reg, data);
>  		if (rc < 0)
>  			return rc;
>  	}

[Severity: High]
Can the new callback return a value large enough to cause an out-of-bounds
write here?

If _pmbus_read_block_data() delegates to the new read_block_data callback,
and that callback returns a size of 34 (I2C_SMBUS_BLOCK_MAX + 2) or greater,
would it cause a buffer overflow?

Just below this hunk in pmbus_debugfs_block_read(), the code does:

    data[rc] = '\n';

Since the size of the data array is 34, an rc of 34 would write past the end
of the array. Is a bounds check needed on rc before using it as an index?

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260728-hwmon-max20826-support-v1-0-224766e0acd1@analog.com?part=3

Re: [PATCH 3/5] hwmon: (pmbus/core) Add mapping function to pmbus_read_block_data()

[Guenter Roeck]

On 7/28/26 09:27, sashiko-bot@kernel.org wrote:
> Thank you for your contribution! Sashiko AI review found 2 potential issue(s) to consider:
> 
> New issues:
> - [High] Stack buffer overflow in `pmbus_debugfs_block_read` due to unbounded return value from custom `read_block_data` callback.
> 
> Pre-existing issues:
> - [High] TOCTOU data race on PMBus hardware page selection during probe.
> --
> 
> --- Patch [3]: [PATCH 3/5] hwmon: (pmbus/core) Add mapping function to pmbus_read_block_data() ---
> commit c4bd4d03203a9fd14a7a2bcff69ec842b441fa68
> Author: Nuno Sá <nuno.sa@analog.com>
> 
> hwmon: (pmbus/core) Add mapping function to pmbus_read_block_data()
> 
> This commit introduces a mapping function to pmbus_read_block_data() in
> preparation for adding support to a device which needs to use its own
> read_block implementation.
> 
>> diff --git a/drivers/hwmon/pmbus/pmbus.h b/drivers/hwmon/pmbus/pmbus.h
>> index 3d5586c67f840..d697939ec8921 100644
>> --- a/drivers/hwmon/pmbus/pmbus.h
>> +++ b/drivers/hwmon/pmbus/pmbus.h
>> @@ -459,6 +459,9 @@ struct pmbus_driver_info {
>>   	int (*read_byte_data)(struct i2c_client *client, int page, int reg);
>>   	int (*read_word_data)(struct i2c_client *client, int page, int phase,
>>   			      int reg);
>> +	/* size of data_buf is I2C_SMBUS_BLOCK_MAX + 2 */
>> +	int (*read_block_data)(struct i2c_client *client, int page, u8 reg,
>> +			       char *data_buf);
> 
> [Severity: High]
> Does this API need a size limit parameter? The comment specifies the buffer
> size is 34, without providing a limit parameter to the callback, which might
> encourage callbacks to return values up to 34.
> 

The maximum amount of data returned must be 32 bytes, not 34. Callers define
a larger buffer size to be able to append, for example, '\n\0'. The function
itself, however, must not return more than 32 bytes of data, plus the data
length in the first byte.

Guenter
> [ ... ]
> 
>> diff --git a/drivers/hwmon/pmbus/pmbus_core.c b/drivers/hwmon/pmbus/pmbus_core.c
>> index 7b58f71985741..ff4572c473b70 100644
>> --- a/drivers/hwmon/pmbus/pmbus_core.c
>> +++ b/drivers/hwmon/pmbus/pmbus_core.c
> [ ... ]
>> @@ -678,7 +698,7 @@ static bool __maybe_unused pmbus_check_block_register(struct i2c_client *client,
>>   	struct pmbus_data *data = i2c_get_clientdata(client);
>>   	char data_buf[I2C_SMBUS_BLOCK_MAX + 2];
>>   
>> -	rv = pmbus_read_block_data(client, page, reg, data_buf);
>> +	rv = _pmbus_read_block_data(client, page, reg, data_buf);
> 
> [Severity: High]
> This is a pre-existing issue, but can this lead to a race condition with
> sysfs accesses during probe?
> 
> Looking at pmbus_init_debugfs(), it traverses pages and reads registers
> without taking the pmbus_lock. This indirectly invokes pmbus_set_page()
> through pmbus_check_block_register():
> 
> pmbus_init_debugfs()
>    pmbus_check_block_register()
>      _pmbus_read_block_data()
>        pmbus_read_block_data()
>          pmbus_set_page()
> 
> If userspace reads or writes to a sysfs attribute concurrently, it acquires
> the lock and sets the hardware page. Could the lockless call from
> pmbus_init_debugfs() interleave between the sysfs thread's pmbus_set_page()
> and its subsequent I2C operation, causing the sysfs thread to hit the wrong
> PMBus page?
> 
>>   	if (rv >= 0 && !(data->flags & PMBUS_SKIP_STATUS_CHECK))
>>   		rv = pmbus_check_status_cml(client);
> [ ... ]
>> @@ -3564,7 +3584,7 @@ static ssize_t pmbus_debugfs_block_read(struct file *file, char __user *buf,
>>   	char data[I2C_SMBUS_BLOCK_MAX + 2] = { 0 };
>>   
>>   	scoped_guard(pmbus_lock, client) {
>> -		rc = pmbus_read_block_data(client, entry->page, entry->reg, data);
>> +		rc = _pmbus_read_block_data(client, entry->page, entry->reg, data);
>>   		if (rc < 0)
>>   			return rc;
>>   	}
> 
> [Severity: High]
> Can the new callback return a value large enough to cause an out-of-bounds
> write here?
> 
> If _pmbus_read_block_data() delegates to the new read_block_data callback,
> and that callback returns a size of 34 (I2C_SMBUS_BLOCK_MAX + 2) or greater,
> would it cause a buffer overflow?
> 
> Just below this hunk in pmbus_debugfs_block_read(), the code does:
> 
>      data[rc] = '\n';
> 
> Since the size of the data array is 34, an rc of 34 would write past the end
> of the array. Is a bounds check needed on rc before using it as an index?
> 

[PATCH 4/5] dt-bindings: hwmon/pmbus: Document MAX20826 and similar devices

[Nuno Sá via B4 Relay]

From: Nuno Sá <nuno.sa@analog.com>

The MAX20826 IC and similar provide a high-density, flexible and scalable
dual-loop solution for high current cores for AI applications. These are
dual loop solutions multiphase voltage regulators. Between Rails A and B,
MAX20855B and MAX20908 supports up to 8 phases total configurable from
8+0 to 4+4 phases, MAX20912 supports up to 12 phases from 12+0 to 6+6,
and MAX20826 and MAX20916 supports up to 16 phases from 16+0 to 8+8.

Co-developed-by: Alexis Czezar Torreno <alexisczezar.torreno@analog.com>
Signed-off-by: Alexis Czezar Torreno <alexisczezar.torreno@analog.com>
Signed-off-by: Nuno Sá <nuno.sa@analog.com>
---
 .../bindings/hwmon/pmbus/adi,max20826.yaml         | 56 ++++++++++++++++++++++
 MAINTAINERS                                        |  7 +++
 2 files changed, 63 insertions(+)

diff --git a/Documentation/devicetree/bindings/hwmon/pmbus/adi,max20826.yaml b/Documentation/devicetree/bindings/hwmon/pmbus/adi,max20826.yaml
new file mode 100644
index 000000000000..6d280f8bf56b
--- /dev/null
+++ b/Documentation/devicetree/bindings/hwmon/pmbus/adi,max20826.yaml
@@ -0,0 +1,56 @@
+# SPDX-License-Identifier: (GPL-2.0-only OR BSD-2-Clause)
+%YAML 1.2
+---
+$id: http://devicetree.org/schemas/hwmon/pmbus/adi,max20826.yaml#
+$schema: http://devicetree.org/meta-schemas/core.yaml#
+
+title: Analog Devices MAX20826 PMBus Voltage Regulator
+
+maintainers:
+  - Nuno Sá <nuno.sa@analog.com>
+
+description:
+  Analog Devices MAX20826 and similar devices are Dual-Output Multiphase
+  High-Current Controller.
+
+properties:
+  compatible:
+    enum:
+      - adi,max20826
+      - adi,max20855b
+      - adi,max20908
+      - adi,max20912
+      - adi,max20916
+
+  reg:
+    maxItems: 1
+
+  avren-gpios:
+    description: GPIO pin to enable/disable the output voltage on Rail A.
+    maxItems: 1
+
+  bvren-gpios:
+    description: GPIO pin to enable/disable the output voltage on Rail B.
+    maxItems: 1
+
+required:
+  - compatible
+  - reg
+
+additionalProperties: false
+
+examples:
+  - |
+    #include <dt-bindings/gpio/gpio.h>
+    i2c {
+        #address-cells = <1>;
+        #size-cells = <0>;
+
+        regulator@64 {
+            compatible = "adi,max20826";
+            reg = <0x64>;
+
+            avren-gpios = <&gpio 0 GPIO_ACTIVE_HIGH>;
+            bvren-gpios = <&gpio 1 GPIO_ACTIVE_HIGH>;
+        };
+    };
diff --git a/MAINTAINERS b/MAINTAINERS
index 716acfc3d7c1..fe8808f3ee4c 100644
--- a/MAINTAINERS
+++ b/MAINTAINERS
@@ -15837,6 +15837,13 @@ F:	Documentation/devicetree/bindings/hwmon/pmbus/adi,max17616.yaml
 F:	Documentation/hwmon/max17616.rst
 F:	drivers/hwmon/pmbus/max17616.c
 
+MAX20826 HARDWARE MONITOR CONTROLLER DRIVER
+M:	Nuno Sá <nuno.sa@analog.com>
+L:	linux-hwmon@vger.kernel.org
+S:	Supported
+W:	https://ez.analog.com/linux-software-drivers
+F:	Documentation/devicetree/bindings/hwmon/pmbus/adi,max20826.yaml
+
 MAX20830 HARDWARE MONITOR DRIVER
 M:	Alexis Czezar Torreno <alexisczezar.torreno@analog.com>
 L:	linux-hwmon@vger.kernel.org

-- 
2.55.0

Re: [PATCH 4/5] dt-bindings: hwmon/pmbus: Document MAX20826 and similar devices

[sashiko-bot]

Thank you for your contribution! Sashiko AI review found 2 potential issue(s) to consider:
- [High] Missing Regulator Subnodes for Dual Rails
- [High] Missing Power Supply Properties
--

--- Patch [4]: [PATCH 4/5] dt-bindings: hwmon/pmbus: Document MAX20826 and similar devices ---
commit 2b08f009cb35e4f4f84d9a87d92b4afdc5d9b230
Author: Nuno Sá <nuno.sa@analog.com>

dt-bindings: hwmon/pmbus: Document MAX20826 and similar devices

This commit introduces device tree bindings for the Analog Devices MAX20826
and similar dual-loop multiphase high-current PMBus voltage regulators. It
defines the schema for the compatible strings, registers, and enable GPIOs.

> diff --git a/Documentation/devicetree/bindings/hwmon/pmbus/adi,max20826.yaml b/Documentation/devicetree/bindings/hwmon/pmbus/adi,max20826.yaml
> new file mode 100644
> index 0000000000000..6d280f8bf56bd
> --- /dev/null
> +++ b/Documentation/devicetree/bindings/hwmon/pmbus/adi,max20826.yaml

[ ... ]

> +  reg:
> +    maxItems: 1
> +
> +  avren-gpios:
> +    description: GPIO pin to enable/disable the output voltage on Rail A.
> +    maxItems: 1
> +
> +  bvren-gpios:
> +    description: GPIO pin to enable/disable the output voltage on Rail B.
> +    maxItems: 1

[Severity: High]
Should this binding use standard regulator child subnodes instead of defining
custom avren-gpios and bvren-gpios at the parent level?

Because additionalProperties is set to false below, standard regulator
subnodes are rejected. This prevents consumers from defining standard
regulator constraints like regulator-min-microvolt and bypasses the
built-in pmbus core support for parsing standard regulators.

Would it be better to model this with standard regulator child nodes using
enable-gpios?

> +
> +required:
> +  - compatible
> +  - reg
> +
> +additionalProperties: false

[Severity: High]
Does this binding need to document standard input power supply properties?

Since additionalProperties is false, board device trees are prohibited from
specifying upstream parent supplies (like vin-supply or vcc-supply). This
could prevent the regulator framework from modeling the power tree, which
might break power sequencing if the parent supply needs to be enabled
before this controller operates.

Should standard input supply properties be permitted here?

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260728-hwmon-max20826-support-v1-0-224766e0acd1@analog.com?part=4

[PATCH 5/5] hwmon: (pmbus) add support for MAX20826 and similar devices

[Nuno Sá via B4 Relay]

From: Nuno Sá <nuno.sa@analog.com>

The MAX20826 IC and similar provide a high-density, flexible and scalable
dual-loop solution for high current cores for AI applications. These are
dual loop solutions multiphase voltage regulators. Between Rails A and B,
MAX20855B and MAX20908 supports up to 8 phases total configurable from
8+0 to 4+4 phases, MAX20912 supports up to 12 phases from 12+0 to 6+6,
and MAX20826 and MAX20916 supports up to 16 phases from 16+0 to 8+8.

Co-developed-by: Alexis Czezar Torreno <alexisczezar.torreno@analog.com>
Signed-off-by: Alexis Czezar Torreno <alexisczezar.torreno@analog.com>
Signed-off-by: Nuno Sá <nuno.sa@analog.com>
---
 Documentation/hwmon/index.rst    |    1 +
 Documentation/hwmon/max20826.rst |  124 +++++
 MAINTAINERS                      |    2 +
 drivers/hwmon/pmbus/Kconfig      |   19 +
 drivers/hwmon/pmbus/Makefile     |    1 +
 drivers/hwmon/pmbus/max20826.c   | 1037 ++++++++++++++++++++++++++++++++++++++
 6 files changed, 1184 insertions(+)

diff --git a/Documentation/hwmon/index.rst b/Documentation/hwmon/index.rst
index 29130df44d12..c63992b292c1 100644
--- a/Documentation/hwmon/index.rst
+++ b/Documentation/hwmon/index.rst
@@ -164,6 +164,7 @@ Hardware Monitoring Kernel Drivers
    max197
    max20730
    max20751
+   max20826
    max20830
    max20860a
    max31722
diff --git a/Documentation/hwmon/max20826.rst b/Documentation/hwmon/max20826.rst
new file mode 100644
index 000000000000..be97312dc612
--- /dev/null
+++ b/Documentation/hwmon/max20826.rst
@@ -0,0 +1,124 @@
+.. SPDX-License-Identifier: GPL-2.0
+
+Kernel driver max20826
+======================
+
+Supported chips:
+
+  * Analog Devices MAX20826
+
+    Prefix: 'max20826'
+
+    Addresses scanned: -
+
+    Datasheet: Datasheet is not publicly available.
+
+  * Analog Devices MAX20855B
+
+    Prefix: 'max20855b'
+
+    Addresses scanned: -
+
+    Datasheet: Datasheet is not publicly available.
+
+  * Analog Devices MAX20908
+
+    Prefix: 'max20908'
+
+    Addresses scanned: -
+
+    Datasheet: Datasheet is not publicly available.
+
+  * Analog Devices MAX20912
+
+    Prefix: 'max20912'
+
+    Addresses scanned: -
+
+    Datasheet: Datasheet is not publicly available.
+
+  * Analog Devices MAX20916
+
+    Prefix: 'max20916'
+
+    Addresses scanned: -
+
+    Datasheet: Datasheet is not publicly available.
+
+Author:
+
+  - Nuno Sá <nuno.sa@analog.com>
+
+
+Description
+-----------
+
+This driver supports hardware monitoring for Analog Devices MAX20826,
+MAX20855B, MAX20908, MAX20912, and MAX20916 multiphase voltage regulator
+controllers with PMBus interface.
+
+The devices are dual-loop, multiphase controllers. Depending on the device and
+configuration, the high-speed processor voltage-control interface can be
+Nvidia PWMVID, Intel SVID, AMD SVI3, or AVSBus. PMBus is used for monitoring,
+configuration, status, and fault reporting.
+
+The driver detects whether the device uses PMBus page mode or direct address
+mode. In direct address mode, rail B is accessed at the rail A I2C address plus
+one. If rail B is present, the driver exposes a second PMBus page.
+
+The driver detects the active number of phases and exposes per-phase input and
+output current attributes through the PMBus virtual phase support.
+
+Usage Notes
+-----------
+
+This driver does not auto-detect devices. You will have to instantiate the
+devices explicitly. Please see Documentation/i2c/instantiating-devices.rst for
+details.
+
+The optional ``avren`` and ``bvren`` GPIOs may be provided to control the rail A
+and rail B hardware enable pins. If regulator support is enabled, the rails are
+also registered through the PMBus regulator framework.
+
+Sysfs entries
+-------------
+
+The following attributes are supported. Limits, alarms, and per-phase entries
+are exposed depending on device capabilities, rail configuration, and detected
+phase count.
+
+=========================== ================================================
+in1_label                   "vin"
+in1_input                   Measured input voltage
+in1_alarm                   Input voltage alarm
+in[2-3]_label               "vout[1-2]"
+in[2-3]_input               Measured output voltage
+in[2-3]_alarm               Output voltage alarm
+in[2-3]_high_speed_en       Enable high-speed voltage-control interface
+in_high_speed_bus           Active high-speed voltage-control interface
+currX_label                 "iinN", "iinN.P", "ioutN", or "ioutN.P"
+currX_input                 Measured input/output current
+currX_alarm                 Current alarm
+powerX_label                "pinN" or "poutN"
+powerX_input                Measured input/output power
+temp[1-2]_input             Measured temperature
+temp[1-2]_alarm             Temperature alarm
+=========================== ================================================
+
+Notes
+-----
+
+``N`` is the rail number, starting at 1. ``P`` is the phase number, starting at
+0. The exact ``currX`` indices depend on the number of present rails and on the
+detected phase count.
+
+``in[2-3]_high_speed_en`` is a per-rail read/write attribute matching the
+standard hwmon voltage channels for the output rails. ``in2_high_speed_en``
+controls rail A / ``vout1``. ``in3_high_speed_en`` controls rail B / ``vout2``
+and is only present if rail B is detected. Writing 1 selects the high-speed
+voltage-control interface for that rail; writing 0 selects PMBus voltage
+control.
+
+``in_high_speed_bus`` is a read-only attribute reporting the high-speed
+interface used by the device, for example ``Nvidia PWMVID``, ``Intel SVID``,
+``AMD SVI3``, or ``AVSBus``.
diff --git a/MAINTAINERS b/MAINTAINERS
index fe8808f3ee4c..8518d55800e9 100644
--- a/MAINTAINERS
+++ b/MAINTAINERS
@@ -15843,6 +15843,8 @@ L:	linux-hwmon@vger.kernel.org
 S:	Supported
 W:	https://ez.analog.com/linux-software-drivers
 F:	Documentation/devicetree/bindings/hwmon/pmbus/adi,max20826.yaml
+F:	Documentation/hwmon/max20826.rst
+F:	drivers/hwmon/pmbus/max20826.c
 
 MAX20830 HARDWARE MONITOR DRIVER
 M:	Alexis Czezar Torreno <alexisczezar.torreno@analog.com>
diff --git a/drivers/hwmon/pmbus/Kconfig b/drivers/hwmon/pmbus/Kconfig
index c8cda160b5f8..8d3568f7347d 100644
--- a/drivers/hwmon/pmbus/Kconfig
+++ b/drivers/hwmon/pmbus/Kconfig
@@ -393,6 +393,25 @@ config SENSORS_MAX20751
 	  This driver can also be built as a module. If so, the module will
 	  be called max20751.
 
+config SENSORS_MAX20826
+	tristate "Analog Devices MAX20826 and similar devices"
+	help
+	  If you say yes here you get hardware monitoring support for Analog
+	  Devices MAX20826 and similar devices.
+
+	  This driver can also be built as a module. If so, the module will
+	  be called max20826.
+
+config SENSORS_MAX20826_REGULATOR
+	bool "Regulator support for Analog Devices MAX20826 and similar devices"
+	depends on SENSORS_MAX20826 && REGULATOR
+	help
+	  If you say yes here you get regulator support for Analog Devices
+	  MAX20826 and similar sensors.
+
+	  This enables the MAX20826 to be used as a regulator device,
+	  providing voltage control through the regulator framework.
+
 config SENSORS_MAX20830
 	tristate "Analog Devices MAX20830"
 	help
diff --git a/drivers/hwmon/pmbus/Makefile b/drivers/hwmon/pmbus/Makefile
index ffc05f493213..15601443470b 100644
--- a/drivers/hwmon/pmbus/Makefile
+++ b/drivers/hwmon/pmbus/Makefile
@@ -38,6 +38,7 @@ obj-$(CONFIG_SENSORS_MAX16601)	+= max16601.o
 obj-$(CONFIG_SENSORS_MAX17616)	+= max17616.o
 obj-$(CONFIG_SENSORS_MAX20730)	+= max20730.o
 obj-$(CONFIG_SENSORS_MAX20751)	+= max20751.o
+obj-$(CONFIG_SENSORS_MAX20826)	+= max20826.o
 obj-$(CONFIG_SENSORS_MAX20830)	+= max20830.o
 obj-$(CONFIG_SENSORS_MAX20860A)	+= max20860a.o
 obj-$(CONFIG_SENSORS_MAX31785)	+= max31785.o
diff --git a/drivers/hwmon/pmbus/max20826.c b/drivers/hwmon/pmbus/max20826.c
new file mode 100644
index 000000000000..1b481cdd8c93
--- /dev/null
+++ b/drivers/hwmon/pmbus/max20826.c
@@ -0,0 +1,1037 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * Hardware monitoring driver for Analog Devices MAX20826 PMBus device
+ *
+ * Copyright 2026 Analog Devices Inc.
+ */
+#include <linux/bitfield.h>
+#include <linux/bitops.h>
+#include <linux/bits.h>
+#include <linux/cleanup.h>
+#include <linux/device.h>
+#include <linux/err.h>
+#include <linux/hwmon-sysfs.h>
+#include <linux/i2c.h>
+#include <linux/gpio/consumer.h>
+#include <linux/module.h>
+#include <linux/mod_devicetable.h>
+#include <linux/pmbus.h>
+#include <linux/property.h>
+#include <linux/regulator/driver.h>
+#include <linux/slab.h>
+#include <linux/sysfs.h>
+#include <linux/string.h>
+#include <linux/types.h>
+#include <linux/unaligned.h>
+
+#include "pmbus.h"
+
+#define MAX20826_REG_RAIL_PHASE_CFG	0xB1
+#define MAX20826_REG_CTRL_MISC		0xCB
+#define MAX20826_OPL_EN_MSK		BIT(7)
+
+#define MAX20826_REG_VOUT_RES		0xDB
+#define MAX20826_VOUT_RES_MSK		BIT(7)
+
+#define MAX20855B_REG_VOUT_VRM		0xD1
+#define MAX20855B_REG_VOUT_VRM_MASK	BIT(4)
+
+#define MAX20826_REG_C_MODEAB		0xDC
+/* Byte 1 bits 7 and 6*/
+#define MAX20826_C_MODEAB_MASK		GENMASK(15, 14)
+
+#define MAX20826_REG_ADDR_MODE		0xEC
+#define MAX20826_PAGE_MODE_MSK		BIT(7)
+#define MAX20826_DIRECT_ADDR_MSK	GENMASK(6, 0)
+
+#define MAX20826_REG_OVERRIDE		0xED
+#define MAX20826_OVERRIDE_MASK		BIT(7)
+
+#define MAX20826_REG_PHASE_DETECT	0xF3
+#define MAX20826_REG_PHASE_READ		0xF4
+
+#define MAX20826_REG_STATUS_MON		0xF9
+#define MAX20826_PHASES_NUM_MASK	GENMASK(7, 3)
+
+#define MAX20826_MAX_PAGES		2
+#define MAX20826_MAX_PHASES		16
+#define MAX20826_PHASES_PER_PAGE	8
+#define MAX20826_INTF_PWMVID		1
+#define MAX20826_INTF_AVSBUS		3
+
+#define MAX20855B_PHASES_NUM_MASK	GENMASK(7, 4)
+#define MAX20855B_MAX_PHASES		8
+#define MAX20908_MAX_PHASES		8
+#define MAX20912_MAX_PHASES		12
+#define MAX20916_MAX_PHASES		16
+
+struct max20826_chip_info {
+	const char *vendor_bus_name;
+	u8 max_phases;
+	unsigned int phase_num_mask;
+	u8 start_index_iin;
+	u8 start_index_iout;
+	bool is_reg_addr_mode_block;
+	bool is_vout_direct;
+	bool select_vrm;
+	bool has_avsbus;
+	bool has_opl;
+	u8 (*count_phases)(const u8 *config, int page);
+};
+
+struct max20826 {
+	const struct max20826_chip_info *chip_info;
+	struct pmbus_driver_info info;
+	struct i2c_client *client;
+	/* RAIL-B direct mode */
+	struct i2c_client *client_b;
+	struct i2c_client *curr_client;
+	struct gpio_desc *avren;
+	struct gpio_desc *bvren;
+	bool vendor_bus;
+	bool high_speed[MAX20826_MAX_PAGES];
+	bool on_off_ctrl[MAX20826_MAX_PAGES];
+	bool opl_enabled[MAX20826_MAX_PAGES];
+};
+
+static u8 __max20826_count_phases(const u8 *config, int page)
+{
+	if (page)
+		return hweight8(config[4]);
+
+	return hweight8(config[0]) + hweight8(config[1]) -
+	       hweight8(config[4]);
+}
+
+static u8 __max20855b_count_phases(const u8 *config, int page)
+{
+	if (page)
+		return hweight8(config[3] & 0x3F);
+
+	return hweight8(config[0]) + hweight8(config[1] & 0x0F) -
+	       hweight8(config[3] & 0x3F);
+}
+
+static u8 __max20908_count_phases(const u8 *config, int page)
+{
+	if (page)
+		return hweight8(config[4]);
+
+	return hweight8(config[0]) + hweight8(config[1] & 0xF0) -
+	       hweight8(config[4]);
+}
+
+static u8 __max20912_count_phases(const u8 *config, int page)
+{
+	if (page)
+		return hweight8(config[4]);
+
+	return hweight8(config[0]) + hweight8(config[1] & 0xFC) -
+	       hweight8(config[4]);
+}
+
+#define to_max20826(p)	container_of(p, struct max20826, info)
+
+enum {
+	RAIL_A,
+	RAIL_B,
+};
+
+static const struct regulator_desc __maybe_unused max20826_reg_desc[] = {
+	PMBUS_REGULATOR("vout", 0),
+	PMBUS_REGULATOR("vout", 1),
+};
+
+static struct i2c_client *max20826_select_rail(struct max20826 *st,
+					       int page, bool probing)
+{
+	int ret;
+
+	/*
+	 * If in direct mode and we want RAIL_B (page 1) just return client_b.
+	 * Otherwise, set the proper page (if page mode) and return RAIL_A.
+	 */
+	if (st->client_b) {
+		/* if 0xff just return the last client */
+		if (page < 0)
+			return st->curr_client;
+		if (page)
+			st->curr_client = st->client_b;
+		else
+			st->curr_client = st->client;
+
+		return st->curr_client;
+	}
+
+	if (!probing)
+		ret = pmbus_set_page(st->client, page, 0xff);
+	else
+		ret = i2c_smbus_write_byte_data(st->client, PMBUS_PAGE, page);
+	if (ret < 0)
+		return ERR_PTR(ret);
+
+	return st->client;
+}
+
+static int max20826_update_byte_data_unsafe(const struct i2c_client *client,
+					    int reg, u8 mask, u8 value)
+{
+	int ret;
+
+	ret = i2c_smbus_read_byte_data(client, reg);
+	if (ret < 0)
+		return ret;
+
+	value = (ret & ~mask) | (value & mask);
+
+	return i2c_smbus_write_byte_data(client, reg, value);
+}
+
+static int max20826_read_block_data_unsafe(const struct i2c_client *client,
+					   int reg, u8 *block, u8 size)
+{
+	u8 block_data[I2C_SMBUS_BLOCK_MAX];
+	int ret;
+
+	ret = i2c_smbus_read_i2c_block_data(client, reg, size + 1,
+					    block_data);
+	if (ret < 0)
+		return ret;
+	if (ret < size + 1)
+		return -EIO;
+
+	/* byte 0 comes with the block length, discard it */
+	memcpy(block, block_data + 1, size);
+	return size;
+}
+
+static int __max20826_read_block_data(struct max20826 *st, int page,
+				      int reg, u8 *block, u8 size)
+{
+	struct i2c_client *rail;
+
+	rail = max20826_select_rail(st, page, false);
+	if (IS_ERR(rail))
+		return PTR_ERR(rail);
+
+	return max20826_read_block_data_unsafe(rail, reg, block, size);
+}
+
+static int __max20826_read_byte_data(struct max20826 *st, int page, int reg)
+{
+	struct i2c_client *rail;
+
+	rail = max20826_select_rail(st, page, false);
+	if (IS_ERR(rail))
+		return PTR_ERR(rail);
+
+	return i2c_smbus_read_byte_data(rail, reg);
+}
+
+static int __max20826_write_byte_data(struct max20826 *st, int page, int reg,
+				      u8 value)
+{
+	struct i2c_client *rail;
+
+	rail = max20826_select_rail(st, page, false);
+	if (IS_ERR(rail))
+		return PTR_ERR(rail);
+
+	return i2c_smbus_write_byte_data(rail, reg, value);
+}
+
+static int __max20826_read_word_data(struct max20826 *st, int page, int reg)
+{
+	struct i2c_client *rail;
+
+	rail = max20826_select_rail(st, page, false);
+	if (IS_ERR(rail))
+		return PTR_ERR(rail);
+
+	return i2c_smbus_read_word_data(rail, reg);
+}
+
+static ssize_t max20826_high_speed_en_show(struct device *dev,
+					   struct device_attribute *devattr,
+					   char *buf)
+{
+	struct i2c_client *client = to_i2c_client(dev->parent);
+	const struct pmbus_driver_info *info = pmbus_get_driver_info(client);
+	unsigned int page = to_sensor_dev_attr(devattr)->index;
+	struct max20826 *st = to_max20826(info);
+
+	return sysfs_emit(buf, "%u\n", st->high_speed[page]);
+}
+
+static ssize_t max20826_high_speed_en_store(struct device *dev,
+					    struct device_attribute *devattr,
+					    const char *buf, size_t count)
+{
+	struct i2c_client *client = to_i2c_client(dev->parent);
+	const struct pmbus_driver_info *info = pmbus_get_driver_info(client);
+	unsigned int page = to_sensor_dev_attr(devattr)->index;
+	struct max20826 *st = to_max20826(info);
+	struct i2c_client *rail;
+	bool high_speed;
+	int ret;
+
+	ret = kstrtobool(buf, &high_speed);
+	if (ret)
+		return ret;
+
+	guard(pmbus_lock)(client);
+
+	rail = max20826_select_rail(st, page, false);
+	if (IS_ERR(rail))
+		return PTR_ERR(rail);
+
+	if (!high_speed) {
+		if (!st->vendor_bus && st->chip_info->has_avsbus) {
+			ret = max20826_update_byte_data_unsafe(rail,
+							       PMBUS_OPERATION,
+							       PB_OPERATION_CONTROL_V_SRC, 0);
+			if (ret)
+				return ret;
+
+			goto out_success;
+		}
+
+		ret = max20826_update_byte_data_unsafe(rail,
+						       MAX20826_REG_OVERRIDE,
+						       MAX20826_OVERRIDE_MASK,
+						       FIELD_PREP(MAX20826_OVERRIDE_MASK, 1));
+		if (ret)
+			return ret;
+
+		goto out_success;
+	}
+
+	if (!st->vendor_bus) {
+		ret = max20826_update_byte_data_unsafe(rail, PMBUS_OPERATION,
+						       PB_OPERATION_CONTROL_V_SRC,
+						       FIELD_PREP(PB_OPERATION_CONTROL_V_SRC, 3));
+		if (ret)
+			return ret;
+
+		goto out_success;
+	}
+
+	ret = max20826_update_byte_data_unsafe(rail, MAX20826_REG_OVERRIDE,
+					       MAX20826_OVERRIDE_MASK, 0);
+	if (ret)
+		return ret;
+
+out_success:
+	st->high_speed[page] = high_speed;
+	return count;
+}
+
+static ssize_t max20826_high_speed_bus_show(struct device *dev,
+					    struct device_attribute *devattr,
+					    char *buf)
+{
+	struct i2c_client *client = to_i2c_client(dev->parent);
+	const struct pmbus_driver_info *info = pmbus_get_driver_info(client);
+	struct max20826 *st = to_max20826(info);
+
+	if (st->vendor_bus)
+		return sysfs_emit(buf, "%s\n", st->chip_info->vendor_bus_name);
+
+	return sysfs_emit(buf, "AVSBus\n");
+}
+
+static SENSOR_DEVICE_ATTR_RW(in2_high_speed_en, max20826_high_speed_en, 0);
+static SENSOR_DEVICE_ATTR_RW(in3_high_speed_en, max20826_high_speed_en, 1);
+static SENSOR_DEVICE_ATTR_RO(in_high_speed_bus, max20826_high_speed_bus, 0);
+
+static int max20826_read_curr(struct max20826 *st, int page, int phase, int reg)
+{
+	u8 val_out[6], start_byte;
+	struct i2c_client *rail;
+	int ret;
+
+	if (phase == 0xff)
+		return __max20826_read_word_data(st, page, reg);
+
+	/*
+	 * On Rail_A phases are ascending (from 1) while on RAIL_B they
+	 * are descending (from 16).
+	 */
+	if (!page)
+		phase += 1;
+	else
+		phase = st->chip_info->max_phases - phase;
+
+	/* phase detect and read are only available on RAIL_A */
+	rail = max20826_select_rail(st, RAIL_A, false);
+	if (IS_ERR(rail))
+		return PTR_ERR(rail);
+
+	ret = i2c_smbus_write_byte_data(rail, MAX20826_REG_PHASE_DETECT, phase);
+	if (ret < 0)
+		return ret;
+
+	/*
+	 * For MAX20826: Byte 2:3 is for phase IOUT 4:5 is for phase IIN.
+	 * For other chips: Byte 0:1 is for phase IOUT 2:3 is for phase IIN.
+	 */
+	if (reg == PMBUS_READ_IIN)
+		start_byte = st->chip_info->start_index_iin;
+	else
+		start_byte = st->chip_info->start_index_iout;
+
+	ret = max20826_read_block_data_unsafe(rail, MAX20826_REG_PHASE_READ,
+					      val_out, sizeof(val_out));
+	if (ret < 0)
+		return ret;
+
+	return get_unaligned_le16(&val_out[start_byte]);
+}
+
+static int max20826_read_word_data(struct i2c_client *client, int page,
+				   int phase, int reg)
+{
+	const struct pmbus_driver_info *info = pmbus_get_driver_info(client);
+	struct max20826 *st = to_max20826(info);
+
+	switch (reg) {
+	case PMBUS_READ_IIN:
+	case PMBUS_READ_IOUT:
+		return max20826_read_curr(st, page, phase, reg);
+	case PMBUS_IOUT_OC_FAULT_LIMIT:
+		if (!st->chip_info->has_opl || !st->opl_enabled[page])
+			return __max20826_read_word_data(st, page, reg);
+		return -EIO;
+	case PMBUS_POUT_OP_FAULT_LIMIT:
+		/*
+		 * If Over Power Limit is enabled, PMBUS_IOUT_OC_FAULT_LIMIT
+		 * shows the power limit and hence we need to report it
+		 * properly in Watts.
+		 */
+		if (!st->chip_info->has_opl)
+			return __max20826_read_word_data(st, page, reg);
+
+		if (st->opl_enabled[page])
+			return __max20826_read_word_data(st, page,
+							 PMBUS_IOUT_OC_FAULT_LIMIT);
+		return -EIO;
+	default:
+		if (reg >= PMBUS_VIRT_BASE)
+			return -EOPNOTSUPP;
+		return __max20826_read_word_data(st, page, reg);
+	}
+}
+
+static int max20826_write_word_data(struct i2c_client *client, int page,
+				    int reg, u16 word)
+{
+	const struct pmbus_driver_info *info = pmbus_get_driver_info(client);
+	struct max20826 *st = to_max20826(info);
+	struct i2c_client *rail;
+
+	rail = max20826_select_rail(st, page, false);
+	if (IS_ERR(rail))
+		return PTR_ERR(rail);
+
+	return i2c_smbus_write_word_data(rail, reg, word);
+}
+
+static int max20826_regulator_enable(struct max20826 *st, int page, u8 byte)
+{
+	struct i2c_client *rail;
+
+	rail = max20826_select_rail(st, page, false);
+	if (IS_ERR(rail))
+		return PTR_ERR(rail);
+
+	/*
+	 * If AVSBus is enabled (bits 5 and 4 set) the device refuses to set bit
+	 * 7 of the OPERATION register. Hence, to workaround this, we first
+	 * clear the bits and then set them all together.
+	 */
+	if (st->chip_info->has_avsbus && !st->vendor_bus && st->high_speed[page] &&
+	    PB_OPERATION_CONTROL_ON & byte) {
+		u8 __byte = byte & ~(PB_OPERATION_CONTROL_V_SRC | PB_OPERATION_CONTROL_ON);
+		int ret;
+
+		ret = i2c_smbus_write_byte_data(rail, PMBUS_OPERATION, __byte);
+		if (ret < 0)
+			return ret;
+	}
+
+	if (page)
+		gpiod_set_value_cansleep(st->bvren, !!(PB_OPERATION_CONTROL_ON & byte));
+	else
+		gpiod_set_value_cansleep(st->avren, !!(PB_OPERATION_CONTROL_ON & byte));
+
+	return i2c_smbus_write_byte_data(rail, PMBUS_OPERATION, byte);
+}
+
+static int max20826_write_byte_data(struct i2c_client *client, int page,
+				    int reg, u8 byte)
+{
+	const struct pmbus_driver_info *info = pmbus_get_driver_info(client);
+	struct max20826 *st = to_max20826(info);
+
+	switch (reg) {
+	case PMBUS_OPERATION:
+		return max20826_regulator_enable(st, page, byte);
+	default:
+		return __max20826_write_byte_data(st, page, reg, byte);
+	}
+}
+
+static int max20826_write_byte(struct i2c_client *client, int page, u8 byte)
+{
+	const struct pmbus_driver_info *info = pmbus_get_driver_info(client);
+	struct max20826 *st = to_max20826(info);
+	struct i2c_client *rail;
+
+	rail = max20826_select_rail(st, page, false);
+	if (IS_ERR(rail))
+		return PTR_ERR(rail);
+
+	return i2c_smbus_write_byte(rail, byte);
+}
+
+static int max20826_regulator_enabled(struct max20826 *st, int page)
+{
+	struct gpio_desc *gpio = page ? st->bvren : st->avren;
+	int on, ret;
+
+	if (gpio) {
+		on = gpiod_get_value_cansleep(gpio);
+		if (on < 0)
+			return on;
+	} else {
+		/* If the gpios are not given, just assume it's on */
+		on = 1;
+	}
+
+	ret = __max20826_read_byte_data(st, page, PMBUS_OPERATION);
+	if (ret < 0)
+		return ret;
+
+	if (st->on_off_ctrl[page])
+		on = (PB_OPERATION_CONTROL_ON & ret) && on;
+
+	ret &= ~PB_OPERATION_CONTROL_ON;
+	return ret | FIELD_PREP(PB_OPERATION_CONTROL_ON, on);
+}
+
+static int max20826_iout_status(struct max20826 *st, int page)
+{
+	int status;
+
+	status = __max20826_read_byte_data(st, page, PMBUS_STATUS_IOUT);
+	if (status < 0 || !st->chip_info->has_opl)
+		return status;
+
+	/*
+	 * If Over power limit is on, the fault condition is still set on the OC
+	 * bit
+	 */
+	if (st->opl_enabled[page] && (status & PB_IOUT_OC_FAULT))
+		return status | PB_POUT_OP_FAULT;
+
+	return status;
+}
+
+static int max20826_read_byte_data(struct i2c_client *client, int page, int reg)
+{
+	const struct pmbus_driver_info *info = pmbus_get_driver_info(client);
+	struct max20826 *st = to_max20826(info);
+
+	switch (reg) {
+	case PMBUS_OPERATION:
+		return max20826_regulator_enabled(st, page);
+	case PMBUS_STATUS_IOUT:
+		return max20826_iout_status(st, page);
+	default:
+		return __max20826_read_byte_data(st, page, reg);
+	}
+}
+
+static int max20826_read_block_data(struct i2c_client *client, int page, u8 reg,
+				    char *data_buf)
+{
+	const struct pmbus_driver_info *info = pmbus_get_driver_info(client);
+	struct max20826 *st = to_max20826(info);
+
+	switch (reg) {
+	case PMBUS_MFR_ID:
+	case PMBUS_MFR_MODEL:
+		return __max20826_read_block_data(st, page, reg, data_buf, 16);
+	case PMBUS_MFR_REVISION:
+		return __max20826_read_block_data(st, page, reg, data_buf, 2);
+	case PMBUS_MFR_DATE:
+		return __max20826_read_block_data(st, page, reg, data_buf, 8);
+	default:
+		return -EOPNOTSUPP;
+	}
+}
+
+static struct pmbus_driver_info max20826_default_info = {
+	.pages = 1,
+	.func[0] = PMBUS_HAVE_VIN | PMBUS_HAVE_VOUT | PMBUS_HAVE_STATUS_VOUT |
+		   PMBUS_HAVE_IIN | PMBUS_HAVE_IOUT | PMBUS_HAVE_STATUS_IOUT |
+		   PMBUS_HAVE_TEMP | PMBUS_HAVE_STATUS_TEMP |
+		   PMBUS_HAVE_STATUS_INPUT | PMBUS_HAVE_PIN | PMBUS_HAVE_POUT |
+		   PMBUS_PHASE_VIRTUAL,
+	.format[PSC_VOLTAGE_IN] = linear,
+	.format[PSC_VOLTAGE_OUT] = direct,
+	.format[PSC_TEMPERATURE] = linear,
+	.format[PSC_CURRENT_IN] = linear,
+	.format[PSC_CURRENT_OUT] = linear,
+	.format[PSC_POWER] = linear,
+	.m[PSC_VOLTAGE_OUT] = 2,
+	.R[PSC_VOLTAGE_OUT] = 3,
+	.write_byte_data = max20826_write_byte_data,
+	.write_byte = max20826_write_byte,
+	.read_byte_data = max20826_read_byte_data,
+	.read_word_data = max20826_read_word_data,
+	.write_word_data = max20826_write_word_data,
+	.read_block_data = max20826_read_block_data,
+#if IS_ENABLED(CONFIG_SENSORS_MAX20826_REGULATOR)
+	.num_regulators = 1,
+	.reg_desc = max20826_reg_desc,
+#endif
+};
+
+static struct attribute *max20826_attrs[MAX20826_MAX_PAGES + 1] = {
+	&sensor_dev_attr_in2_high_speed_en.dev_attr.attr,
+	&sensor_dev_attr_in3_high_speed_en.dev_attr.attr,
+	&sensor_dev_attr_in_high_speed_bus.dev_attr.attr,
+};
+
+static int max20826_detect_addr_mode(struct max20826 *st)
+{
+	struct device *dev = &st->client->dev;
+	int ret, val;
+	u8 val_buf[3];
+
+	/*
+	 * The idea is that after a POR, we are in page0 in which case we can
+	 * read MAX20826_REG_ADDR_MODE. If we fail to read
+	 * MAX20826_REG_ADDR_MODE, it might be due to a soft reset or unbinding
+	 * the device in which case the device could be left in page 1. Hence,
+	 * let's just try to change to page 0 and error out if we can't.
+	 *
+	 * OTOH, if the addressing mode is direct, we should be able to
+	 * read MAX20826_REG_ADDR_MODE. If not, we'll fail setting the page.
+	 *
+	 * REG_ADDR_MODE, MAX20855B has 6 bytes, other devices have 1. Byte 3 is
+	 * the equivalent of the single byte of the others.
+	 */
+	if (st->chip_info->is_reg_addr_mode_block)
+		val = max20826_read_block_data_unsafe(st->client,
+						      MAX20826_REG_ADDR_MODE,
+						      val_buf, sizeof(val_buf));
+	else
+		val = i2c_smbus_read_byte_data(st->client, MAX20826_REG_ADDR_MODE);
+
+	if (val < 0) {
+		ret = i2c_smbus_write_byte_data(st->client, PMBUS_PAGE, 0);
+		if (ret < 0)
+			return dev_err_probe(dev, ret,
+					     "Failed to change to page 0\n");
+
+		/* try again! */
+		if (st->chip_info->is_reg_addr_mode_block)
+			val = max20826_read_block_data_unsafe(st->client,
+							      MAX20826_REG_ADDR_MODE,
+							      val_buf, sizeof(val_buf));
+		else
+			val = i2c_smbus_read_byte_data(st->client,
+						       MAX20826_REG_ADDR_MODE);
+		if (val < 0)
+			return dev_err_probe(dev, val,
+					     "Failed to read MAX20826_REG_ADDR_MODE\n");
+	}
+
+	if (st->chip_info->is_reg_addr_mode_block)
+		val = val_buf[2];
+
+	if (val & MAX20826_PAGE_MODE_MSK)
+		return 0;
+
+	/*
+	 * If in direct mode, rail b will be accessible from the addr of RAIL_A
+	 * +1
+	 */
+	st->client_b = devm_i2c_new_dummy_device(dev, st->client->adapter,
+						 st->client->addr + 1);
+	if (IS_ERR(st->client_b))
+		return PTR_ERR(st->client_b);
+
+	return 0;
+}
+
+static int max20826_detect_phases(struct max20826 *st, struct i2c_client *rail,
+				  struct pmbus_driver_info *info, u8 page,
+				  u8 expected)
+{
+	int ret, ret2 = 0, oper_save = -1;
+	u8 status_mon[3];
+	u8 n_phases;
+
+	ret = i2c_smbus_read_byte_data(rail, PMBUS_OPERATION);
+	if (ret < 0)
+		return ret;
+
+	/*
+	 * In order to detect the number of phases, we need to be regulating.
+	 * It is also assumed we're already in the page we want to detect the
+	 * phases.
+	 */
+	if (!(ret & PB_OPERATION_CONTROL_ON)) {
+		oper_save = ret;
+		ret |= PB_OPERATION_CONTROL_ON;
+		ret = i2c_smbus_write_byte_data(rail, PMBUS_OPERATION, ret);
+		if (ret < 0)
+			return ret;
+	}
+
+	ret = max20826_read_block_data_unsafe(rail, MAX20826_REG_STATUS_MON,
+					      status_mon, sizeof(status_mon));
+	if (ret < 0)
+		goto out_restore_oper;
+
+	n_phases = field_get(st->chip_info->phase_num_mask, status_mon[1]);
+	if (n_phases != expected) {
+		ret = dev_err_probe(&st->client->dev, -EIO,
+				    "Number of phases mismatch: expected=%u, detected=%u\n",
+				    expected, n_phases);
+		goto out_restore_oper;
+	}
+
+	info->phases[page] = n_phases;
+	for (unsigned int phase = 0; phase < n_phases; phase++)
+		info->pfunc[phase] = PMBUS_HAVE_IOUT | PMBUS_HAVE_IIN;
+
+out_restore_oper:
+	if (oper_save >= 0)
+		ret2 = i2c_smbus_write_byte_data(rail, PMBUS_OPERATION,
+						 oper_save);
+
+	return ret < 0 ? ret : ret2;
+}
+
+static int max20826_get_rail_config(struct max20826 *st,
+				    struct i2c_client *rail, u8 page)
+{
+	u8 ctrl_misc[2];
+	int ret;
+
+	/*
+	 * Check if we need to control PMBUS_OPERATION in addition to the
+	 * CONTROL pin.
+	 */
+	ret = i2c_smbus_read_byte_data(rail, PMBUS_ON_OFF_CONFIG);
+	if (ret < 0)
+		return ret;
+
+	st->on_off_ctrl[page] = !!(ret & PB_ON_OFF_CONFIG_OPERATION_REQ);
+
+	/*
+	 * See if Over Power Limit is enabled. This will impact how
+	 * OC_FAULT_LIMIT is handled.
+	 */
+	if (st->chip_info->has_opl) {
+		ret = max20826_read_block_data_unsafe(rail, MAX20826_REG_CTRL_MISC,
+						      ctrl_misc, sizeof(ctrl_misc));
+		if (ret < 0)
+			return ret;
+
+		st->opl_enabled[page] = !!(ctrl_misc[0] & MAX20826_OPL_EN_MSK);
+	}
+
+	if (st->vendor_bus) {
+		ret = i2c_smbus_read_byte_data(rail, MAX20826_REG_OVERRIDE);
+		if (ret < 0)
+			return ret;
+
+		st->high_speed[page] = !(ret & MAX20826_OVERRIDE_MASK);
+		return 0;
+	}
+
+	ret = i2c_smbus_read_byte_data(rail, PMBUS_OPERATION);
+	if (ret < 0)
+		return ret;
+
+	if (FIELD_GET(PB_OPERATION_CONTROL_V_SRC, ret) == MAX20826_INTF_AVSBUS)
+		st->high_speed[page] = true;
+
+	return 0;
+}
+
+static int max20826_add_attrs(struct device *dev,
+			      struct pmbus_driver_info *info)
+{
+	unsigned int attr, last = ARRAY_SIZE(max20826_attrs) - 1;
+	struct attribute_group *group;
+
+	info->groups = devm_kcalloc(dev, 2, sizeof(void *), GFP_KERNEL);
+	if (!info->groups)
+		return -ENOMEM;
+
+	group = devm_kzalloc(dev, sizeof(*group), GFP_KERNEL);
+	if (!group)
+		return -ENOMEM;
+
+	info->groups[0] = group;
+
+	group->attrs = devm_kcalloc(dev, info->pages + 2, sizeof(void *),
+				    GFP_KERNEL);
+	if (!group->attrs)
+		return -ENOMEM;
+
+	for (attr = 0; attr < info->pages; attr++)
+		group->attrs[attr] = max20826_attrs[attr];
+
+	group->attrs[attr] = max20826_attrs[last];
+
+	return 0;
+}
+
+static int max20826_setup_vout_format(struct max20826 *st)
+{
+	struct pmbus_driver_info *info = &st->info;
+	int ret;
+
+	if (st->chip_info->is_vout_direct) {
+		u8 vout_res[2];
+
+		ret = __max20826_read_block_data(st, RAIL_A,
+						 MAX20826_REG_VOUT_RES,
+						 vout_res, sizeof(vout_res));
+		if (ret < 0)
+			return ret;
+
+		/* check for 1 mv/LSB */
+		if (MAX20826_VOUT_RES_MSK & vout_res[1])
+			info->m[PSC_VOLTAGE_OUT] = 1;
+
+		return 0;
+	}
+
+	st->info.format[PSC_VOLTAGE_OUT] = vid;
+	for (unsigned int page = 0; page < info->pages; page++) {
+		if (!st->chip_info->select_vrm) {
+			/* only vr12 in this case */
+			st->info.vrm_version[page] = vr12;
+			continue;
+		}
+
+		ret = __max20826_read_byte_data(st, page,
+						MAX20855B_REG_VOUT_VRM);
+		if (ret < 0)
+			return ret;
+
+		if (ret & MAX20855B_REG_VOUT_VRM_MASK)
+			st->info.vrm_version[page] = vr13;
+		else
+			st->info.vrm_version[page] = vr12;
+	}
+
+	return 0;
+}
+
+static int max20826_setup(struct max20826 *st)
+{
+	struct pmbus_driver_info *info = &st->info;
+	u8 config[5], expected_phases;
+	struct device *dev = &st->client->dev;
+	struct i2c_client *rail;
+	int ret;
+
+	ret = max20826_detect_addr_mode(st);
+	if (ret < 0)
+		return ret;
+
+	if (st->chip_info->has_avsbus) {
+		/*
+		 * After max20826_detect_addr_mode() we can just use RAIL_A client for
+		 * the following operations as even if in page mode, we must be in
+		 * page 0 by now. We also just need to check this for chips that might
+		 * also support selecting between custom or AVSBus.
+		 */
+		ret = i2c_smbus_read_word_data(st->client, MAX20826_REG_C_MODEAB);
+		if (ret < 0)
+			return ret;
+
+		if (FIELD_GET(MAX20826_C_MODEAB_MASK, ret) == MAX20826_INTF_PWMVID)
+			st->vendor_bus = true;
+	} else {
+		st->vendor_bus = true;
+	}
+
+	ret = max20826_read_block_data_unsafe(st->client,
+					      MAX20826_REG_RAIL_PHASE_CFG,
+					      config, sizeof(config));
+	if (ret < 0)
+		return ret;
+
+	/*
+	 * If we have phases in RAIL_B, we need to subtract them on config[1] as
+	 * those bits are also set in case RAIL_B has phases.
+	 */
+	expected_phases = st->chip_info->count_phases(config, RAIL_A);
+	ret = max20826_detect_phases(st, st->client, info, RAIL_A,
+				     expected_phases);
+	if (ret < 0)
+		return ret;
+
+	ret = max20826_get_rail_config(st, st->client, RAIL_A);
+	if (ret < 0)
+		return ret;
+
+	/* Let's see if RAIL_B is present */
+	rail = max20826_select_rail(st, RAIL_B, true);
+	if (IS_ERR(rail))
+		return PTR_ERR(rail);
+
+	/* Let's see if there's something on RAIL_B */
+	st->bvren = devm_gpiod_get_optional(dev, "bvren", GPIOD_OUT_HIGH);
+	if (IS_ERR(st->bvren))
+		return PTR_ERR(st->bvren);
+
+	expected_phases = st->chip_info->count_phases(config, RAIL_B);
+	ret = max20826_detect_phases(st, rail, info, RAIL_B, expected_phases);
+	if (ret < 0)
+		return ret;
+
+	if (info->phases[RAIL_B]) {
+		info->pages = MAX20826_MAX_PAGES;
+		info->func[RAIL_B] = PMBUS_HAVE_VOUT | PMBUS_HAVE_STATUS_VOUT |
+				     PMBUS_HAVE_IIN | PMBUS_HAVE_IOUT | PMBUS_HAVE_STATUS_IOUT |
+				     PMBUS_HAVE_TEMP | PMBUS_HAVE_STATUS_TEMP |
+				     PMBUS_HAVE_STATUS_INPUT | PMBUS_HAVE_PIN | PMBUS_HAVE_POUT |
+				     PMBUS_PHASE_VIRTUAL;
+
+		if (IS_ENABLED(CONFIG_SENSORS_MAX20826_REGULATOR))
+			info->num_regulators = 2;
+
+		ret = max20826_get_rail_config(st, rail, RAIL_B);
+		if (ret < 0)
+			return ret;
+	}
+
+	ret = max20826_setup_vout_format(st);
+	if (ret < 0)
+		return ret;
+
+	return max20826_add_attrs(&st->client->dev, info);
+}
+
+static int max20826_probe(struct i2c_client *client)
+{
+	struct max20826 *st;
+	int ret;
+
+	st = devm_kzalloc(&client->dev, sizeof(*st), GFP_KERNEL);
+	if (!st)
+		return -ENOMEM;
+
+	st->client = client;
+	memcpy(&st->info, &max20826_default_info, sizeof(st->info));
+
+	st->chip_info = i2c_get_match_data(client);
+	if (!st->chip_info)
+		return -EINVAL;
+
+	st->avren = devm_gpiod_get_optional(&client->dev, "avren", GPIOD_OUT_HIGH);
+	if (IS_ERR(st->avren))
+		return PTR_ERR(st->avren);
+
+	ret = max20826_setup(st);
+	if (ret)
+		return ret;
+
+	return pmbus_do_probe(client, &st->info);
+}
+
+static const struct max20826_chip_info chip_info_max20826 = {
+	.vendor_bus_name = "Nvidia PWMVID",
+	.max_phases = MAX20826_MAX_PHASES,
+	.phase_num_mask = MAX20826_PHASES_NUM_MASK,
+	.start_index_iin = 4,
+	.start_index_iout = 2,
+	.is_vout_direct = true,
+	.has_avsbus = true,
+	.has_opl = true,
+	.count_phases = __max20826_count_phases,
+};
+
+static const struct max20826_chip_info chip_info_max20855b = {
+	.vendor_bus_name = "Intel SVID",
+	.max_phases = MAX20855B_MAX_PHASES,
+	.phase_num_mask = MAX20855B_PHASES_NUM_MASK,
+	.start_index_iin = 2,
+	.start_index_iout = 0,
+	.is_reg_addr_mode_block = true,
+	.select_vrm = true,
+	.count_phases = __max20855b_count_phases,
+};
+
+static const struct max20826_chip_info chip_info_max20908 = {
+	.vendor_bus_name = "AMD SVI3",
+	.max_phases = MAX20908_MAX_PHASES,
+	.phase_num_mask = MAX20826_PHASES_NUM_MASK,
+	.start_index_iin = 2,
+	.start_index_iout = 0,
+	.count_phases = __max20908_count_phases,
+};
+
+static const struct max20826_chip_info chip_info_max20912 = {
+	.vendor_bus_name = "AMD SVI3",
+	.max_phases = MAX20912_MAX_PHASES,
+	.phase_num_mask = MAX20826_PHASES_NUM_MASK,
+	.start_index_iin = 2,
+	.start_index_iout = 0,
+	.count_phases = __max20912_count_phases,
+};
+
+static const struct max20826_chip_info chip_info_max20916 = {
+	.vendor_bus_name = "AMD SVI3",
+	.max_phases = MAX20916_MAX_PHASES,
+	.phase_num_mask = MAX20826_PHASES_NUM_MASK,
+	.start_index_iin = 2,
+	.start_index_iout = 0,
+	.count_phases = __max20826_count_phases,
+};
+
+static const struct i2c_device_id max20826_id[] = {
+	{ "max20826", (kernel_ulong_t)&chip_info_max20826 },
+	{ "max20855b", (kernel_ulong_t)&chip_info_max20855b },
+	{ "max20908", (kernel_ulong_t)&chip_info_max20908 },
+	{ "max20912", (kernel_ulong_t)&chip_info_max20912 },
+	{ "max20916", (kernel_ulong_t)&chip_info_max20916 },
+	{ }
+};
+MODULE_DEVICE_TABLE(i2c, max20826_id);
+
+static const struct of_device_id max20826_of_match[] = {
+	{ .compatible = "adi,max20826", .data = &chip_info_max20826 },
+	{ .compatible = "adi,max20855b", .data = &chip_info_max20855b },
+	{ .compatible = "adi,max20908", .data = &chip_info_max20908 },
+	{ .compatible = "adi,max20912", .data = &chip_info_max20912 },
+	{ .compatible = "adi,max20916", .data = &chip_info_max20916 },
+	{ }
+};
+MODULE_DEVICE_TABLE(of, max20826_of_match);
+
+static struct i2c_driver max20826_driver = {
+	.driver = {
+		.name = "max20826",
+		.of_match_table = max20826_of_match,
+	},
+	.probe = max20826_probe,
+	.id_table = max20826_id,
+};
+module_i2c_driver(max20826_driver);
+
+MODULE_AUTHOR("Nuno Sá <nuno.sa@analog.com>");
+MODULE_DESCRIPTION("PMBus driver for MAX20826");
+MODULE_LICENSE("GPL");
+MODULE_IMPORT_NS("PMBUS");

-- 
2.55.0

Re: [PATCH 5/5] hwmon: (pmbus) add support for MAX20826 and similar devices

[sashiko-bot]

Thank you for your contribution! Sashiko AI review found 6 potential issue(s) to consider:
- [Critical] Fatal NULL pointer dereference during probe for devices configured in page mode due to premature `pmbus_set_page()` calls.
- [High] The driver forcefully turns on voltage regulator outputs and unconditionally asserts hardware enable GPIOs during probe, bypassing system power sequencing.
- [High] Probe fails on AVSBus-enabled systems due to a missing register workaround during phase detection.
- [Medium] Inconsistent register redirection in custom PMBus callbacks breaks power limit configuration when Over Power Limit (OPL) is enabled.
- [Medium] Global PMBus commands in direct addressing mode are arbitrarily routed to a non-deterministic I2C client.
- [Medium] Hardware state (GPIO) is modified without rollback if the subsequent PMBus I2C operation fails.
--

--- Patch [5]: [PATCH 5/5] hwmon: (pmbus) add support for MAX20826 and similar devices ---
commit 8b8b70e16f48ffac7a09e30045ceff6976a47f0b
Author: Nuno Sá <nuno.sa@analog.com>
Subject: hwmon: (pmbus) add support for MAX20826 and similar devices

The MAX20826 IC and similar provide a high-density, flexible and scalable
dual-loop solution for high current cores for AI applications. These are
dual loop solutions multiphase voltage regulators. Between Rails A and B,
MAX20855B and MAX20908 supports up to 8 phases total configurable from
8+0 to 4+4 phases, MAX20912 supports up to 12 phases from 12+0 to 6+6,
and MAX20826 and MAX20916 supports up to 16 phases from 16+0 to 8+8.

> diff --git a/drivers/hwmon/pmbus/max20826.c b/drivers/hwmon/pmbus/max20826.c
> new file mode 100644
> index 0000000000000..1b481cdd8c930
> --- /dev/null
> +++ b/drivers/hwmon/pmbus/max20826.c
> @@ -156,12 +156,12 @@ static struct i2c_client *max20826_select_rail(struct max20826 *st,
>  	if (st->client_b) {
>  		/* if 0xff just return the last client */
>  		if (page < 0)
>  			return st->curr_client;

[Severity: Medium]
Will global commands be arbitrarily routed to a non-deterministic I2C client?

For global commands, the PMBus core passes page = -1. When page < 0, this
simply returns st->curr_client without updating it, which holds whichever
rail was accessed last. If userspace reads PMBUS_STATUS_CML globally, won't
it only query the CML status of the most recently accessed rail, potentially
masking communication faults on the other rail? Similarly, will global write
protection apply to only one rail?

>  		if (page)
>  			st->curr_client = st->client_b;
>  		else
>  			st->curr_client = st->client;
>  
>  		return st->curr_client;
>  	}
>  
>  	if (!probing)
>  		ret = pmbus_set_page(st->client, page, 0xff);

[Severity: Critical]
Does this cause a fatal NULL pointer dereference during probe for devices
configured in page mode?

During probe, max20826_setup() -> max20826_setup_vout_format() calls
__max20826_read_block_data() (as seen below). The wrapper hardcodes
probing = false. Because of this flag, max20826_select_rail() calls
pmbus_set_page(). 

pmbus_set_page() internally retrieves pmbus_data via i2c_get_clientdata().
However, because pmbus_do_probe() has not been executed yet, clientdata is
NULL, leading to a fatal dereference of data->info when pmbus_set_page()
is called.

>  	else
>  		ret = i2c_smbus_write_byte_data(st->client, PMBUS_PAGE, page);

[ ... ]

> @@ -201,11 +201,11 @@ static int __max20826_read_block_data(struct max20826 *st, int page,
>  	struct i2c_client *rail;
>  
>  	rail = max20826_select_rail(st, page, false);
>  	if (IS_ERR(rail))
>  		return PTR_ERR(rail);

[ ... ]

> @@ -396,17 +396,28 @@ static int max20826_read_word_data(struct i2c_client *client, int page,
[ ... ]
>  	case PMBUS_POUT_OP_FAULT_LIMIT:
>  		/*
>  		 * If Over Power Limit is enabled, PMBUS_IOUT_OC_FAULT_LIMIT
>  		 * shows the power limit and hence we need to report it
>  		 * properly in Watts.
>  		 */
>  		if (!st->chip_info->has_opl)
>  			return __max20826_read_word_data(st, page, reg);
>  
>  		if (st->opl_enabled[page])
>  			return __max20826_read_word_data(st, page,
>  							 PMBUS_IOUT_OC_FAULT_LIMIT);
>  		return -EIO;

[ ... ]

> @@ -417,14 +428,14 @@ static int max20826_write_word_data(struct i2c_client *client, int page,
>  				    int reg, u16 word)
>  {
>  	const struct pmbus_driver_info *info = pmbus_get_driver_info(client);
>  	struct max20826 *st = to_max20826(info);
>  	struct i2c_client *rail;
>  
>  	rail = max20826_select_rail(st, page, false);
>  	if (IS_ERR(rail))
>  		return PTR_ERR(rail);
>  
>  	return i2c_smbus_write_word_data(rail, reg, word);
>  }

[Severity: Medium]
Will writes to the power limit target the wrong PMBus register directly on
the hardware when Over Power Limit (OPL) is enabled?

While max20826_read_word_data() properly implements the redirection mapping
from PMBUS_POUT_OP_FAULT_LIMIT to PMBUS_IOUT_OC_FAULT_LIMIT when OPL is
enabled, max20826_write_word_data() blindly passes all writes straight to
the wire without interception. Should the redirection logic also be applied
to the write callback?

[ ... ]

> @@ -451,10 +462,10 @@ static int max20826_regulator_enable(struct max20826 *st, int page, u8 byte)
>  	if (page)
>  		gpiod_set_value_cansleep(st->bvren, !!(PB_OPERATION_CONTROL_ON & byte));
>  	else
>  		gpiod_set_value_cansleep(st->avren, !!(PB_OPERATION_CONTROL_ON & byte));
>  
>  	return i2c_smbus_write_byte_data(rail, PMBUS_OPERATION, byte);
>  }

[Severity: Medium]
Can the hardware state (GPIO) become desynchronized with the software PMBus
state if the subsequent I2C operation fails?

If i2c_smbus_write_byte_data() returns a negative error code, the regulator
core treats the operation as failed, but the hardware GPIO pin was left in
the newly toggled state without being rolled back.

[ ... ]

> @@ -807,11 +807,11 @@ static int max20826_detect_phases(struct max20826 *st, struct i2c_client *rail,
>  	if (!(ret & PB_OPERATION_CONTROL_ON)) {
>  		oper_save = ret;
>  		ret |= PB_OPERATION_CONTROL_ON;
>  		ret = i2c_smbus_write_byte_data(rail, PMBUS_OPERATION, ret);

[Severity: High]
Does this unconditionally force the PMBus rail ON during phase detection,
potentially bypassing system power sequencing rules?

In max20826_probe() and max20826_setup() (as seen below), the driver
unconditionally requests the optional avren and bvren GPIOs with
GPIOD_OUT_HIGH. Then, here in max20826_detect_phases(), it forces the
PMBus OPERATION register to ON just to read the active phase count. While
it restores the PMBus register state, the GPIOs are left permanently asserted
high. Could this violate strict SoC power sequencing rules and damage the
processor or downstream loads?

Also, as documented in max20826_regulator_enable(), if AVSBus is active
(bits 4 and 5 set), the device hardware refuses to set bit 7 unless the
AVSBus bits are first cleared. Because max20826_detect_phases() fails to
apply this workaround before the I2C write, won't the device ignore the ON
command, resulting in a read of 0 phases and aborting the probe on
AVSBus-enabled systems?

>  		if (ret < 0)
>  			return ret;
>  	}

[ ... ]

> @@ -898,11 +898,11 @@ static int max20826_setup_vout_format(struct max20826 *st)
>  	if (st->chip_info->is_vout_direct) {
>  		u8 vout_res[2];
>  
>  		ret = __max20826_read_block_data(st, RAIL_A,
>  						 MAX20826_REG_VOUT_RES,
>  						 vout_res, sizeof(vout_res));

[ ... ]

> @@ -946,8 +946,8 @@ static int max20826_setup(struct max20826 *st)
>  	/* Let's see if there's something on RAIL_B */
>  	st->bvren = devm_gpiod_get_optional(dev, "bvren", GPIOD_OUT_HIGH);
>  	if (IS_ERR(st->bvren))
>  		return PTR_ERR(st->bvren);

[ ... ]

> @@ -999,8 +999,8 @@ static int max20826_probe(struct i2c_client *client)
>  	st->chip_info = i2c_get_match_data(client);
>  	if (!st->chip_info)
>  		return -EINVAL;
>  
>  	st->avren = devm_gpiod_get_optional(&client->dev, "avren", GPIOD_OUT_HIGH);
>  	if (IS_ERR(st->avren))
>  		return PTR_ERR(st->avren);

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260728-hwmon-max20826-support-v1-0-224766e0acd1@analog.com?part=5