Re: [PATCH V13 7/9] regulator: Add a regulator driver for the PM8008 PMIC

[Satya Priya Kakitapalli (Temp)]

On 5/28/2022 2:32 AM, Stephen Boyd wrote:
> HTML mail? Please send plaintext next time.
>
> Quoting Satya Priya Kakitapalli (Temp) (2022-05-27 01:24:19)
> > On 5/21/2022 8:26 AM, Stephen Boyd wrote:
> >
> >      Quoting Satya Priya (2022-05-20 03:49:35)
> >
> >          diff --git a/drivers/regulator/qcom-pm8008-regulator.c b/drivers/regulator/qcom-pm8008-regulator.c
> >          new file mode 100644
> >          index 0000000..6e815c6
> >          --- /dev/null
> >          +++ b/drivers/regulator/qcom-pm8008-regulator.c
> >          @@ -0,0 +1,225 @@
> >          +// SPDX-License-Identifier: GPL-2.0-only
> >          +/* Copyright (c) 2022, The Linux Foundation. All rights reserved. */
> >          +
> >          +#include <linux/device.h>
> >          +#include <linux/kernel.h>
> >          +#include <linux/mfd/qcom_pm8008.h>
> >          +#include <linux/module.h>
> >          +#include <linux/of.h>
> >          +#include <linux/of_device.h>
> >
> >      What in of_device.h is used?
> >
> >
> > struct of_device_id
> That struct is defined in mod_devicetable.h, not of_device.h
>
> >          +#include <linux/platform_device.h>
> >          +#include <linux/regmap.h>
> >          +#include <linux/regulator/driver.h>
> >          +
> [...]
> >          +};
> >          +
> >          +static int pm8008_regulator_get_voltage(struct regulator_dev *rdev)
> >          +{
> >          +       struct pm8008_regulator *pm8008_reg = rdev_get_drvdata(rdev);
> >          +
> >          +       return pm8008_reg->voltage_selector;
> >
> >      Can this read the hardware instead of caching the value from
> >      pm8008_regulator_set_voltage()?
> >
> >
> > I can use the regmap_bulk_read like below (which was present in the earlier
> > versions)
> Please do
>
> >          +       }
> >          +
> >          +       pm8008_reg->dev = dev;
> >          +
> >          +       rc = of_property_read_string(dev->of_node, "regulator-name", &name);
> >          +       if (rc)
> >          +               return rc;
> >          +
> >          +       /* get the required regulator data */
> >          +       for (i = 0; i < ARRAY_SIZE(reg_data); i++)
> >          +               if (strstr(name, reg_data[i].name))
> >
> >      Why not find this via reg/address instead? It would save storing the
> >      regulator name in the reg_data table.
> >
> >
> > You mean match this using base address? then we should add base address in the
> > reg_data table. We will need the name to be stored in reg_data table anyway for
> > the pm8008_reg->rdesc.of_match
> Why? Now that this driver binds to each node individually the usage of
> of_match doesn't make any sense to me. Can you set 'struct
> regulator_config::dev' instead and not set of_match?


Currently we are setting regulator_config::dev as dev->parent i.e., 
pm8008@8, because the parent supplies are present under pm8008@8, to get 
the regulators mapped correctly to the parent supplies we are using 
dev->parent.

If we do not set of_match in regulator descriptor, 
regulator_of_get_init_node() would return NULL, causing init_data to be 
NULL during regulator_register and regulators are not getting probed. 
This can be resolved, if we get the init_data during pm8008_probe 
itself. I'll do that in the next version.