On pon, 2014-11-03 at 15:40 +0100, Javier Martinez Canillas wrote:
> Some regulators support their operating mode to be changed on startup
> or by consumers when the system is running while others only support
> their operating mode to be changed while the system has entered in a
> suspend state.
>
> The regulator Device Tree binding documents a set of properties to
> configure the regulators operating modes from a FDT. This patch builds
> on (40e20d6 regulator: of: Add support for parsing regulator_state for
> suspend state) and adds support to parse those properties and fill the
> regulator constraints so the regulator core can call the right suspend
> handlers when the system enters into sleep.
>
> The modes are defined in the Device Tree using the hardware specific
> modes supported by the regulators. Regulator drivers have to define a
> translation function that is used to map the hardware specific modes
> to the standard ones.
>
> Signed-off-by: Javier Martinez Canillas <javier.martinez@xxxxxxxxxxxxxxx>
> ---
>
> Changes in v4:
> - Parse the properties in the core and map using driver provided functions.
> Suggested by Mark Brown
>
> Changes in v3:
> - Use the standard suspend states binding instead of custom properties.
> Suggested by Mark Brown
>
> drivers/regulator/of_regulator.c | 22 ++++++++++++++++++++--
> 1 file changed, 20 insertions(+), 2 deletions(-)
>
> diff --git a/drivers/regulator/of_regulator.c b/drivers/regulator/of_regulator.c
> index cbc1d71..cd65885 100644
> --- a/drivers/regulator/of_regulator.c
> +++ b/drivers/regulator/of_regulator.c
> @@ -25,7 +25,8 @@ static const char *const regulator_states[PM_SUSPEND_MAX + 1] = {
> };
>
> static void of_get_regulation_constraints(struct device_node *np,
> - struct regulator_init_data **init_data)
> + struct regulator_init_data **init_data,
> + const struct regulator_desc *desc)
> {
> const __be32 *min_uV, *max_uV;
> struct regulation_constraints *constraints = &(*init_data)->constraints;
> @@ -81,6 +82,14 @@ static void of_get_regulation_constraints(struct device_node *np,
> if (!ret)
> constraints->enable_time = pval;
>
> + if (!of_property_read_u32(np, "regulator-initial-mode", &pval)) {
> + if (desc && desc->map_modes)
> + constraints->initial_mode = desc->map_modes(pval);
> + else
> + pr_warn("%s: failed to parse regulator-initial-mode\n",
> + np->name);
> + }
> +
Here's a hidden assumption that if driver does not provide map_modes
then any "regulator-initial-mode" property is not valid. Shouldn't this
be mentioned somewhere? Maybe in description of map_modes callback?
Best regards,
Krzysztof
--
To unsubscribe from this list: send the line "unsubscribe devicetree" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
