Message-ID: <CAKew6eVZuOUH8RHQbffCGX7i=6dM-qOJXojoKrOsM3V5ugEjHA@mail.gmail.com>
Date:	Tue, 15 May 2012 19:17:11 +0530
From:	Yadwinder Singh Brar <yadi.brar01@...il.com>
To:	Mark Brown <broonie@...nsource.wolfsonmicro.com>
Cc:	Yadwinder Singh <yadi.brar@...sung.com>,
	linux-samsung-soc@...r.kernel.org, linux-kernel@...r.kernel.org,
	linux-arm-kernel@...ts.infradead.org
Subject: Re: [PATCH 2/2] regulator: Add support for MAX77686.

Hi Mark,

On Thu, May 10, 2012 at 12:17 AM, Mark Brown
<broonie@...nsource.wolfsonmicro.com> wrote:
> On Wed, May 09, 2012 at 09:54:55PM +0530, Yadwinder Singh wrote:
>
>> +/* Voltage maps in mV */
>> +static const struct voltage_map_desc ldo_voltage_map_desc = {
>> +     .min = 800,     .max = 3950,    .step = 50,     .n_bits = 6,
>> +};                           /* LDO3 ~ 5, 9 ~ 14, 16 ~ 26 */
>
> Hrm, funnily enough I was just thinking about factoring this stuff out
> into the core after a conversation with Graeme Gregory the other week.
> Let's do that...
>
>> +     [MAX77686_EN32KHZ_AP] = NULL,
>> +     [MAX77686_EN32KHZ_CP] = NULL,
>
> Now that the generic clock API is in mainline these should be moved over
> to use it.
>
>> +static int max77686_get_voltage_unit(int rid)
>> +{
>> +     int unit = 0;
>> +
>> +     switch (rid) {
>> +     case MAX77686_BUCK2...MAX77686_BUCK4:
>> +             unit = 1;       /* BUCK2,3,4 is uV */
>> +             break;
>> +     default:
>> +             unit = 1000;
>
> Why not just list everything in uV?
>
>> +static int max77686_get_voltage(struct regulator_dev *rdev)
>> +{
>
> Implement get_voltage_sel().
>
>> +static inline int max77686_get_voltage_proper_val(const struct voltage_map_desc
>> +                                               *desc, int min_vol,
>> +                                               int max_vol)
>> +{
>> +     int i = 0;
>> +
>> +     if (desc == NULL)
>> +             return -EINVAL;
>> +
>> +     if (max_vol < desc->min || min_vol > desc->max)
>> +             return -EINVAL;
>> +
>> +     while (desc->min + desc->step * i < min_vol &&
>> +            desc->min + desc->step * i < desc->max)
>> +             i++;
>
> Why are you iterating here?  Calculate!  Though like I say let's factor
> this out anyway.
>
>> +     if (rid == MAX77686_BUCK2 || rid == MAX77686_BUCK3 ||
>> +         rid == MAX77686_BUCK4) {
>> +             /* If the voltage is increasing */
>> +             if (org < i)
>> +                     udelay(DIV_ROUND_UP(desc->step * (i - org),
>> +                                         ramp[max77686->ramp_delay]));
>> +     }
>
> Don't do this, implement set_voltage_time_sel().
>
>> +     .enable = max77686_reg_enable,
>> +     .disable = max77686_reg_disable,
>> +     .set_suspend_enable = max77686_reg_enable,
>> +     .set_suspend_disable = max77686_reg_disable,
>
> You've got the same ops for suspend and non-suspend cases here, this is
> clearly buggy.
>
>> +/* count the number of regulators to be supported in pmic */
>> +     pdata->num_regulators = 0;
>
> Coding style.
>
>> +     if (iodev->dev->of_node) {
>> +             ret = max77686_pmic_dt_parse_pdata(iodev, pdata);
>> +             if (ret)
>> +                     return ret;
>
> This ought to use of_regulator_match().
>

While using it I am seeing that though it reduces few lines of code in
our driver , it adds a huge array(of_regulator_match[])in which we
have to populate the strings(name) which we already have in
regulator_desc[ ], Isn't it overhead ?

>> +     }
>> +
>> +     if (!pdata) {
>> +             dev_err(&pdev->dev, "platform data not found\n");
>> +             return -ENODEV;
>> +     }
>
> This should be totally fine.
>
>> +     max77686 = kzalloc(sizeof(struct max77686_data), GFP_KERNEL);
>> +     if (!max77686)
>> +             return -ENOMEM;
>
> devm_kzalloc().
>
>> +     if (pdata->ramp_delay) {
>> +             max77686->ramp_delay = pdata->ramp_delay;
>> +             max77686_update_reg(i2c, MAX77686_REG_BUCK2CTRL1,
>> +                     RAMP_VALUE, RAMP_MASK);
>
> This appears not to actually use the value passed in as platform_data.
>
>> +
>> +     for (i = 0; i < pdata->num_regulators; i++) {
>> +             const struct voltage_map_desc *desc;
>> +             int id = pdata->regulators[i].id;
>> +
>> +             desc = reg_voltage_map[id];
>> +             if (desc)
>> +                     regulators[id].n_voltages =
>> +                         (desc->max - desc->min) / desc->step + 1;
>> +
>> +             rdev[i] = regulator_register(&regulators[id], max77686->dev,
>> +                                          pdata->regulators[i].initdata,
>> +                                          max77686, NULL);
>
> No, you should unconditionally register all regulators the device
> physically has.  This is useful for debug and simplifies the code.
>

If we have to use only 2 or 3 regulators on our board out off 36 or
lets take a case if our chip supports 50/100 regulators, I think i
will a overhead to register all unused regulators as well as
populating all the regulators in DT or platform data.

Regards,
Yadwinder.
> _______________________________________________
> linux-arm-kernel mailing list
> linux-arm-kernel@...ts.infradead.org
> http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/

Powered by blists - more mailing lists