[PATCH v3] hwmon: (pmbus/core): Implement regulator get_status

[PATCH v3] hwmon: (pmbus/core): Implement regulator get_status

[Naresh Solanki]

From: Patrick Rudolph <patrick.rudolph@9elements.com>

Add get_status for pmbus_regulator_ops.

Changes:
- use lock throughout the function
- Avoid line continuation upto 100 column
- Optimize use of & and | operator
- Check for VOUT, IOUT, TEMPERATURE bit in status word before checking
  respective status register for fault.
- Report regulator current status.

Signed-off-by: Patrick Rudolph <patrick.rudolph@9elements.com>
Signed-off-by: Naresh Solanki <Naresh.Solanki@9elements.com>
---
 drivers/hwmon/pmbus/pmbus_core.c | 88 ++++++++++++++++++++++++++++++++
 1 file changed, 88 insertions(+)

diff --git a/drivers/hwmon/pmbus/pmbus_core.c b/drivers/hwmon/pmbus/pmbus_core.c
index 7ec04934747e..f7a322d5b453 100644
--- a/drivers/hwmon/pmbus/pmbus_core.c
+++ b/drivers/hwmon/pmbus/pmbus_core.c
@@ -2851,6 +2851,93 @@ static int pmbus_regulator_get_error_flags(struct regulator_dev *rdev, unsigned
 	return 0;
 }
 
+static int pmbus_regulator_get_status(struct regulator_dev *rdev)
+{
+	struct device *dev = rdev_get_dev(rdev);
+	struct i2c_client *client = to_i2c_client(dev->parent);
+	struct pmbus_data *data = i2c_get_clientdata(client);
+	u8 page = rdev_get_id(rdev);
+	int status, status2, ret;
+
+	mutex_lock(&data->update_lock);
+	status = pmbus_get_status(client, page, PMBUS_STATUS_WORD);
+	if (status < 0) {
+		ret = status;
+		goto unlock;
+	}
+
+	if (status & PB_STATUS_OFF) {
+		ret = REGULATOR_STATUS_OFF;
+		goto unlock;
+	}
+
+	/* If regulator is ON & reports power good then return ON */
+	if (!(status & PB_STATUS_POWER_GOOD_N)) {
+		ret = REGULATOR_STATUS_ON;
+		goto unlock;
+	}
+
+	if (status & (PB_STATUS_VIN_UV | PB_STATUS_IOUT_OC | PB_STATUS_VOUT_OV |
+	   PB_STATUS_UNKNOWN)){
+		ret = REGULATOR_STATUS_ERROR;
+		goto unlock;
+	}
+
+	if (status & PB_STATUS_VOUT && data->info->func[page] & PMBUS_HAVE_STATUS_VOUT) {
+		status2 = _pmbus_read_byte_data(client, page,
+					 PMBUS_STATUS_VOUT);
+		if (status2 < 0) {
+			ret = status2;
+			goto unlock;
+		}
+		if (status2 & (PB_VOLTAGE_OV_FAULT | PB_VOLTAGE_UV_FAULT)) {
+			ret = REGULATOR_STATUS_ERROR;
+			goto unlock;
+		}
+	}
+
+	if (status & PB_STATUS_IOUT_POUT && data->info->func[page] & PMBUS_HAVE_STATUS_IOUT) {
+		status2 = _pmbus_read_byte_data(client, page, PMBUS_STATUS_IOUT);
+		if (status2 < 0) {
+			ret = status2;
+			goto unlock;
+		}
+		if (status2 & (PB_POUT_OP_FAULT | PB_IOUT_UC_FAULT | PB_IOUT_OC_LV_FAULT |
+		   PB_IOUT_OC_FAULT)) {
+			ret = REGULATOR_STATUS_ERROR;
+			goto unlock;
+		}
+	}
+
+	if (status & PB_STATUS_INPUT && data->info->func[page] & PMBUS_HAVE_STATUS_INPUT) {
+		status2 = _pmbus_read_byte_data(client, page, PMBUS_STATUS_INPUT);
+		if (status2 < 0) {
+			ret = status2;
+			goto unlock;
+		}
+		if (status2 & (PB_IIN_OC_FAULT | PB_VOLTAGE_OV_FAULT | PB_VOLTAGE_UV_FAULT)) {
+			ret = REGULATOR_STATUS_ERROR;
+			goto unlock;
+		}
+	}
+
+	if (status & PB_STATUS_TEMPERATURE && data->info->func[page] & PMBUS_HAVE_STATUS_TEMP) {
+		status2 = _pmbus_read_byte_data(client, page, PMBUS_STATUS_TEMPERATURE);
+		if (status2 < 0) {
+			ret = status2;
+			goto unlock;
+		}
+		if (status2 & (PB_TEMP_UT_FAULT | PB_TEMP_OT_FAULT)) {
+			ret = REGULATOR_STATUS_ERROR;
+			goto unlock;
+		}
+	}
+
+unlock:
+	mutex_unlock(&data->update_lock);
+	return ret;
+}
+
 static int pmbus_regulator_get_low_margin(struct i2c_client *client, int page)
 {
 	struct pmbus_data *data = i2c_get_clientdata(client);
@@ -2991,6 +3078,7 @@ const struct regulator_ops pmbus_regulator_ops = {
 	.disable = pmbus_regulator_disable,
 	.is_enabled = pmbus_regulator_is_enabled,
 	.get_error_flags = pmbus_regulator_get_error_flags,
+	.get_status = pmbus_regulator_get_status,
 	.get_voltage = pmbus_regulator_get_voltage,
 	.set_voltage = pmbus_regulator_set_voltage,
 	.list_voltage = pmbus_regulator_list_voltage,

base-commit: 27fea302952d8c90cafbdbee96bafeca03544401
-- 
2.37.3

Re: [PATCH v3] hwmon: (pmbus/core): Implement regulator get_status

[Guenter Roeck]

On Sat, Nov 19, 2022 at 07:25:16PM +0100, Naresh Solanki wrote:
> From: Patrick Rudolph <patrick.rudolph@9elements.com>
> 
> Add get_status for pmbus_regulator_ops.
> 
> Changes:
> - use lock throughout the function
> - Avoid line continuation upto 100 column
> - Optimize use of & and | operator
> - Check for VOUT, IOUT, TEMPERATURE bit in status word before checking
>   respective status register for fault.
> - Report regulator current status.
> 
Change log should be after '---'

Also, when looking into this, I realized that we already have
pmbus_regulator_get_error_flags() which has somewhat overlapping
functionality. Would it be possible to utilize that function to get
the error status instead of more or less hand-coding it ?

Thanks,
Guenter

Re: [PATCH v3] hwmon: (pmbus/core): Implement regulator get_status

[Naresh Solanki]

Hi Guenter,

On 20-11-2022 07:44 pm, Guenter Roeck wrote:
> On Sat, Nov 19, 2022 at 07:25:16PM +0100, Naresh Solanki wrote:
>> From: Patrick Rudolph <patrick.rudolph@9elements.com>
>>
>> Add get_status for pmbus_regulator_ops.
>>
>> Changes:
>> - use lock throughout the function
>> - Avoid line continuation upto 100 column
>> - Optimize use of & and | operator
>> - Check for VOUT, IOUT, TEMPERATURE bit in status word before checking
>>    respective status register for fault.
>> - Report regulator current status.
>>
> Change log should be after '---'
Sure
> 
> Also, when looking into this, I realized that we already have
> pmbus_regulator_get_error_flags() which has somewhat overlapping
> functionality. Would it be possible to utilize that function to get
> the error status instead of more or less hand-coding it ?
Sure. But looking at get_status scope, I feel that we are already 
checking the following:
1. Check if regulator is off & yes then return off
2. If not off then it should be on & has power good then return on
3. If no power good then this itself is error condition so return error. 
Decoding status register for fault just double confirms that & may not 
be needed in get_status unless we are going to print or utilize the 
error bit information in some other way but that is out of scope of 
get_status call unless I missed something.

Does this approach makes sense ?
> 
> Thanks,
> Guenter