On Fri, May 26, 2017 at 7:38 PM, Mark Brown [off-list ref] wrote:

On Fri, May 26, 2017 at 02:35:16PM +0800, Guodong Xu wrote:

Overall this driver needs quite a lot of modernization, it's at least a
couple of years out of date in how it's using the framework - there's
barely any use of helpers.  It does look like it should be fairly easy
to get it up to date though, it's mostly going to be a case of deleting
code that's reimplementing helpers rather than anything else.

quoted

+/*
+ * struct hi6421c530_regulator_pdata - Hi6421V530 regulator data
+ * of platform device.
+ * @lock: mutex to serialize regulator enable
+ */
+struct hi6421v530_regulator_pdata {
+     struct mutex lock;
+};

This isn't platform data so it probably shouldn't be called pdata.  I
also can't tell what the lock is protecting, every use seems to be a
call to regmap_update_bits() which is atomic anyway - could we just drop
the whole thing?

In original hi6421 chip, this lock can protect from enabling multiple
LDOs simultaneously. Because it cannot afford such surging current.

I will double check whether this is still the case for hi6421v530. If
not, I will remove the whole thing.

quoted

+static int hi6421v530_regulator_enable(struct regulator_dev *rdev)
+{
+     struct hi6421v530_regulator_pdata *pdata;
+     int ret = 0;
+
+     pdata = dev_get_drvdata(rdev->dev.parent);
+     mutex_lock(&pdata->lock);
+
+     ret = regmap_update_bits(rdev->regmap, rdev->desc->enable_reg,
+                     rdev->desc->enable_mask,
+                     1 << (ffs(rdev->desc->enable_mask) - 1));
+
+     mutex_unlock(&pdata->lock);
+     return ret;
+}

This looks like it should be able to use the regmap helpers for all the
enable operations rather than open coding.

quoted

+static int hi6421v530_regulator_set_voltage(struct regulator_dev *rdev,
+                                             unsigned int sel)
+{
+     struct hi6421v530_regulator_pdata *pdata;
+     int ret = 0;
+
+     pdata = dev_get_drvdata(rdev->dev.parent);
+     mutex_lock(&pdata->lock);
+
+     ret = regmap_update_bits(rdev->regmap, rdev->desc->vsel_reg,
+                             rdev->desc->vsel_mask,
+                             sel << (ffs(rdev->desc->vsel_mask) - 1));

Same for all the voltage operations :(

quoted

+     rdev->constraints->valid_modes_mask = info->mode_mask;
+     rdev->constraints->valid_ops_mask |=
+                     REGULATOR_CHANGE_VOLTAGE | REGULATOR_CHANGE_MODE;

The driver should *never* modify constraints, it's up to the machine
integration to say what can be supported on a given board.

quoted

+     np = of_get_child_by_name(dev->parent->of_node, "regulators");
+     if (!np)
+             return -ENODEV;
+
+     ret = of_regulator_match(dev, np,
+                              hi6421v530_regulator_match,
+                              ARRAY_SIZE(hi6421v530_regulator_match));

Don't open code this, use of_match and regulators_node.

I will fix that.

Thanks Mark for your review.

-Guodong
--
To unsubscribe from this list: send the line "unsubscribe devicetree" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html