Re: [PATCH 2/2] U300 AB3100 boardinfo

From: Linus Walleij
Date: Mon Aug 31 2009 - 18:02:19 EST


2009/8/31 Mark Brown <broonie@xxxxxxxxxxxxxxxxxxxxxxxxxxx>:

> On Sun, 2009-08-30 at 23:30 +0200, Linus Walleij wrote:
>
>>  arch/arm/mach-u300/Makefile    |    1 +
>>  arch/arm/mach-u300/i2c.c       |  304 ++++++++++++++++++++++++++++++++++++++++
>>  arch/arm/mach-u300/regulator.c |  150 ++++++++++++++++++++
>
> Is mach-u300 for all machines with a given processor or is it for a
> specific system?

It is a series of systems (U330, U365, U335) but they all feature the same
connections and rails to the AB3100 circuit so regulator-wise they are
identical.

>> +static struct regulator_consumer_supply supply_ldo_c[] = {
>> +     {
>> +             .dev_name = "ab3100-codec",
>> +             .supply = "vaudio", /* Powers the codec */
>> +     },
>> +};
>
> Does the CODEC really take a single supply called vaudio or are there
> several inputs to the CODEC?

It's only one actually.

>> +             /* Buck converter in sleep mode routing and constraints */
>> +             {
>> +                     .constraints = {
>> +                             .min_uV = 0,
>> +                             .max_uV = 1800000,
>> +                             .valid_modes_mask = REGULATOR_MODE_NORMAL,
>> +                             .valid_ops_mask =
>> +                             REGULATOR_CHANGE_VOLTAGE |
>> +                             REGULATOR_CHANGE_STATUS,
>> +                     },
>> +                     .num_consumer_supplies = ARRAY_SIZE(supply_buck_sleep),
>> +                     .consumer_supplies = supply_buck_sleep,
>> +             },
>
> Hrm, I didn't pick this up when reading the regulator driver but does
> this mean that you're registering two different regulators for the same
> buck convertor?

Yeah I did. So I removed the sleep mode "regulators" and added
the set_suspend_voltage() hook instead. Works like a charm.

>> +/*
>> + * Hog the regulators needed to power up the board.
>> + */
>> +static int __init u300_init_boardpower(void)
>> +{
>> +     int err;
>> +     u32 val;
>> +
>> +     pr_info("U300: setting up board power\n");
>> +     radio_power = regulator_get(NULL, "vrad");
>> +     if (IS_ERR(radio_power)) {
>> +             pr_err("could not get vrad\n");
>> +             return PTR_ERR(radio_power);
>> +     }
>> +     err = regulator_enable(radio_power);
>> +     if (err) {
>> +             pr_err("could not enable vrad\n");
>> +             return err;
>> +     }
>
> This all looks like it can be supported with the existing constraints
> code; look at boot_on and always_on. Normally the only thing that should
> need deviceless supplies like you have here is cpufreq (due to the fact
> that there's no struct device available for cpufreq to use).

I switched to using always_on and boot_on and it's just as good.
Though I kept one of them: vana15, because disabling that
regulator is what I use for shutdown hook for the system.

Thanks for all the great comments Mark, v3 is due in a minute.

Yours,
Linus Walleij
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/