Skip to contentSkip to conversation

hwmon: (ucd9000) Add voltage monitor types

PATCH In review Jim Wright 2 participants 4 messages Jun 8, 2022 Hardware monitoring

lore
Share & tools
lore
maintainer
#1 lore raw
In reply to Jim Wright
Quoted 24 lines · Jim Wright

The UCD90240 and UCD90320 devices support monitor types "Input Voltage", "Voltage With AVS", and "Input Voltage With AVS", add support to driver to recognize these types as voltage monitors.

---
  drivers/hwmon/pmbus/ucd9000.c | 6 ++++++
  1 file changed, 6 insertions(+)
drivers/hwmon/pmbus/ucd9000.c +3 -0
index 75fc770c9e40..28cc214d4d6b 100644
@@ -45,6 +45,9 @@ enum chips { ucd9000, ucd90120, ucd90124, ucd90160, ucd90320, ucd9090,
  #define UCD9000_MON_TEMPERATURE	2
  #define UCD9000_MON_CURRENT	3
  #define UCD9000_MON_VOLTAGE_HW	4
+#define UCD9000_MON_INPUT_VOLTAGE	5
+#define UCD9000_MON_VOLTAGE_AVS	6
+#define UCD9000_MON_INPUT_VOLTAGE_AVS	7
 
  #define UCD9000_NUM_FAN		4
 

drivers/hwmon/pmbus/ucd9000.c static int ucd9000_probe(struct i2c_client *client) · line 577

Quoted 9 lines
drivers/hwmon/pmbus/ucd9000.c +3 -0
@@ -566,6 +569,9 @@ static int ucd9000_probe(struct i2c_client *client)
  		switch (UCD9000_MON_TYPE(block_buffer[i])) {
  		case UCD9000_MON_VOLTAGE:
  		case UCD9000_MON_VOLTAGE_HW:
+		case UCD9000_MON_INPUT_VOLTAGE:
+		case UCD9000_MON_VOLTAGE_AVS:
+		case UCD9000_MON_INPUT_VOLTAGE_AVS:
  			info->func[page] |= PMBUS_HAVE_VOUT
  			  | PMBUS_HAVE_STATUS_VOUT;
  			break;

I don't think it makes sense to claim VOUT support if the chip is configured to monitor input voltages. This should probably be something like

...

drivers/hwmon/pmbus/ucd9000.c static int ucd9000_probe(struct i2c_client *client) · line 577

Quoted 4 lines
drivers/hwmon/pmbus/ucd9000.c +1 -0
+		case UCD9000_MON_VOLTAGE_AVS:
  			info->func[page] |= PMBUS_HAVE_VOUT
  			  | PMBUS_HAVE_STATUS_VOUT;
  			break;
		case UCD9000_MON_INPUT_VOLTAGE:
		case UCD9000_MON_INPUT_VOLTAGE_AVS:
			info->func[page] |= PMBUS_HAVE_VIN;

break;

with appropriate mapping code to map the READ_VIN command for the affected pages to READ_VOUT. Question is if the limit registers on those pages are also reporting the limits using the vout limit commands; if so, those should be mapped as well.

Guenter

Jim Wright

author
#0 lore raw

PATCH 1/1 hwmon: (ucd9000) Add voltage monitor types

The UCD90240 and UCD90320 devices support monitor types "Input Voltage", "Voltage With AVS", and "Input Voltage With AVS", add support to driver to recognize these types as voltage monitors.

---
 drivers/hwmon/pmbus/ucd9000.c | 6 ++++++
 1 file changed, 6 insertions(+)
drivers/hwmon/pmbus/ucd9000.c +6 -0
index 75fc770c9e40..28cc214d4d6b 100644
@@ -45,6 +45,9 @@ enum chips { ucd9000, ucd90120, ucd90124, ucd90160, ucd90320, ucd9090,
 #define UCD9000_MON_TEMPERATURE	2
 #define UCD9000_MON_CURRENT	3
 #define UCD9000_MON_VOLTAGE_HW	4
+#define UCD9000_MON_INPUT_VOLTAGE	5
+#define UCD9000_MON_VOLTAGE_AVS	6
+#define UCD9000_MON_INPUT_VOLTAGE_AVS	7
 
 #define UCD9000_NUM_FAN		4
 
@@ -566,6 +569,9 @@ static int ucd9000_probe(struct i2c_client *client)
 		switch (UCD9000_MON_TYPE(block_buffer[i])) {
 		case UCD9000_MON_VOLTAGE:
 		case UCD9000_MON_VOLTAGE_HW:
+		case UCD9000_MON_INPUT_VOLTAGE:
+		case UCD9000_MON_VOLTAGE_AVS:
+		case UCD9000_MON_INPUT_VOLTAGE_AVS:
 			info->func[page] |= PMBUS_HAVE_VOUT
 			  | PMBUS_HAVE_STATUS_VOUT;
 			break;
Signature · 1 line
2.32.0
author
#2 lore raw
In reply to Guenter Roeck
Quoted 15 lines · Guenter Roeck

I don't think it makes sense to claim VOUT support if the chip is configured to monitor input voltages. This should probably be something like

...

Quoted 4 lines

+        case UCD9000_MON_VOLTAGE_AVS:
               info->func[page] |= PMBUS_HAVE_VOUT | PMBUS_HAVE_STATUS_VOUT; break;

        case UCD9000_MON_INPUT_VOLTAGE:
        case UCD9000_MON_INPUT_VOLTAGE_AVS:
            info->func[page] |= PMBUS_HAVE_VIN;
             break;

Quoted 6 lines

with appropriate mapping code to map the READ_VIN command for the affected pages to READ_VOUT. Question is if the limit registers on those pages are also reporting the limits using the vout limit commands; if so, those should be mapped as well.

Guenter

Hi Guenter,

Thank you for the review. I'll drop adding the input voltage types and resend the patch.

Jim Wright

author
#3 lore raw
Quoted 21 lines · Jim Wright
Quoted 20 lines · Guenter Roeck

I don't think it makes sense to claim VOUT support if the chip is configured to monitor input voltages. This should probably be something like

...
 > +        case UCD9000_MON_VOLTAGE_AVS:
 >               info->func[page] |= PMBUS_HAVE_VOUT >                 | PMBUS_HAVE_STATUS_VOUT; >               break;
         case UCD9000_MON_INPUT_VOLTAGE:
         case UCD9000_MON_INPUT_VOLTAGE_AVS:
             info->func[page] |= PMBUS_HAVE_VIN; break;

with appropriate mapping code to map the READ_VIN command for the affected pages to READ_VOUT. Question is if the limit registers on those pages are also reporting the limits using the vout limit commands; if so, those should be mapped as well.

Quoted 7 lines
Quoted 1 line

Guenter

Hi Guenter,

Thank you for the review. I'll drop adding the input voltage types and resend the patch.

Jim Wright

After a second look, it's the input voltage type that I need. Will revise as suggested and resubmit.

Thanks,
Jim Wright