On Wed, Jan 25, 2012 at 04:27:48PM +0530, Laxman Dewangan wrote:

> @@ -450,6 +489,29 @@ static int tps65910_set_mode(struct regulator_dev *dev, 
> unsigned int mode)
>       struct tps65910 *mfd = pmic->mfd;
>       int reg, value, id = rdev_get_id(dev);
>  
> +     /*
> +      * If regulator is controlled through external control then
> +      * mode can be identified by the input level of EN1/EN2/EN3.
> +      * If it is HIGH then regulators is on, full power.
> +      * If it is LOW then:
> +      *       - the regulator is set off if its corresponding Control
> +      *         bit = 0 in SLEEP_KEEP_XXX_ON.
> +      *       - the regulator is set in low-power mode if its corresponding
> +      *         control bit = 1 in SLEEP_KEEP_XXX_ON register.
> +      */

This really isn't what the set_mode() API is for - especially the fact
that it supports turning the regulator off which really isn't what
set_mode() is supposed to do.  A generic driver using this API isn't
going to play too well.

> +     if (pmic->board_ext_control[id]) {
> +             u8 regoffs = (pmic->ext_sleep_control[id] >> 8) & 0xFF;
> +             u8 bit_pos = (1 << pmic->ext_sleep_control[id] & 0xFF);
> +             int ret = 0;
> +             if ((mode == REGULATOR_MODE_IDLE) ||
> +                     (mode == REGULATOR_MODE_STANDBY))
> +                     ret = tps65910_set_bits(mfd,
> +                             TPS65910_SLEEP_KEEP_LDO_ON + regoffs, bit_pos);

As a coding style thing this should be a switch statement.

Attachment: signature.asc
Description: Digital signature

Reply via email to