Re: [PATCH v2 1/3] hwmon: (pmbus/core): Update regulator flag map

From: Naresh Solanki
Date: Tue Nov 29 2022 - 02:55:29 EST


Hi Guenter,

On 29-11-2022 04:11 am, Guenter Roeck wrote:
On 11/28/22 09:47, Naresh Solanki wrote:
Add regulator flag map for PMBUS status byte & status input.

Signed-off-by: Naresh Solanki <Naresh.Solanki@xxxxxxxxxxxxx>

You are adding a lot of input errors here. The regulator documentation
only covers output errors. I am not sure if this set of changes is
really appropriate. You'll have to make a much better case for those changes;
from what I can see they are all controversial and were originally left out
on purpose.
I felt it may be worth to monitor status input, but you feel otherwise then shall I remove this in next revision ?

---
  drivers/hwmon/pmbus/pmbus_core.c | 30 ++++++++++++++++++++++--------
  1 file changed, 22 insertions(+), 8 deletions(-)

diff --git a/drivers/hwmon/pmbus/pmbus_core.c b/drivers/hwmon/pmbus/pmbus_core.c
index 95e95783972a..f5caceaaef2a 100644
--- a/drivers/hwmon/pmbus/pmbus_core.c
+++ b/drivers/hwmon/pmbus/pmbus_core.c
@@ -2752,6 +2752,15 @@ struct pmbus_regulator_status_category {
  static const struct pmbus_regulator_status_category pmbus_regulator_flag_map[] = {
      {
+        .func = -1,

This would need a comment. I don't really see the benefit over the original
code.
I pulled that in so as to handle it in same way as other status register.

+        .reg = PMBUS_STATUS_BYTE,
+        .bits = (const struct pmbus_regulator_status_assoc[]) {
+            { PB_STATUS_IOUT_OC,   REGULATOR_ERROR_OVER_CURRENT },
+            { PB_STATUS_VOUT_OV,   REGULATOR_ERROR_REGULATION_OUT },
+            { PB_STATUS_VIN_UV,    REGULATOR_ERROR_UNDER_VOLTAGE },
+            { },
+        },
+    }, {
          .func = PMBUS_HAVE_STATUS_VOUT,
          .reg = PMBUS_STATUS_VOUT,
          .bits = (const struct pmbus_regulator_status_assoc[]) {
@@ -2768,6 +2777,7 @@ static const struct pmbus_regulator_status_category pmbus_regulator_flag_map[] =
              { PB_IOUT_OC_WARNING, REGULATOR_ERROR_OVER_CURRENT_WARN },
              { PB_IOUT_OC_FAULT,      REGULATOR_ERROR_OVER_CURRENT },
              { PB_IOUT_OC_LV_FAULT,   REGULATOR_ERROR_OVER_CURRENT },
+            { PB_POUT_OP_FAULT,      REGULATOR_ERROR_OVER_CURRENT },

OP_FAULT (power fault) and over current are really not the same thing.

I agree. But thats best I could think of. Not sure if there is better REGULATOR_ERROR_* code for this scenario. Suggestions?
              { },
          },
      }, {
@@ -2778,6 +2788,18 @@ static const struct pmbus_regulator_status_category pmbus_regulator_flag_map[] =
              { PB_TEMP_OT_FAULT,      REGULATOR_ERROR_OVER_TEMP },
              { },
          },
+    }, {
+        .func = PMBUS_HAVE_STATUS_INPUT,
+        .reg = PMBUS_STATUS_INPUT,
+        .bits = (const struct pmbus_regulator_status_assoc[]) {
+            { PB_IIN_OC_FAULT,       REGULATOR_ERROR_OVER_CURRENT },
+            { PB_IIN_OC_WARNING, REGULATOR_ERROR_OVER_CURRENT_WARN },
+            { PB_VOLTAGE_UV_FAULT,   REGULATOR_ERROR_UNDER_VOLTAGE },
+            { PB_VOLTAGE_UV_WARNING, REGULATOR_ERROR_UNDER_VOLTAGE_WARN },
+            { PB_VOLTAGE_OV_WARNING, REGULATOR_ERROR_OVER_VOLTAGE_WARN },
+            { PB_VOLTAGE_OV_FAULT, REGULATOR_ERROR_OVER_VOLTAGE_WARN },

fault -> warning ? Shouldn't this be REGULATOR_ERROR_FAIL (Regulator
output has failed) ?

Yes. REGULATOR_ERROR_FAIL is best fit here. Will update in next revision.
+            { },
+        },
      },
  };
@@ -2834,14 +2856,6 @@ static int pmbus_regulator_get_error_flags(struct regulator_dev *rdev, unsigned
          if (status & PB_STATUS_POWER_GOOD_N)
              *flags |= REGULATOR_ERROR_REGULATION_OUT;
      }
-    /*
-     * Unlike most other status bits, PB_STATUS_{IOUT_OC,VOUT_OV} are
-     * defined strictly as fault indicators (not warnings).
-     */
-    if (status & PB_STATUS_IOUT_OC)
-        *flags |= REGULATOR_ERROR_OVER_CURRENT;
-    if (status & PB_STATUS_VOUT_OV)
-        *flags |= REGULATOR_ERROR_REGULATION_OUT;
      /*
       * If we haven't discovered any thermal faults or warnings via

base-commit: 9494c53e1389b120ba461899207ac8a3aab2632c