Re: [PATCH] mfd: tps65219: Add support for soft shutdown via sys-off API

From: Lee Jones
Date: Thu May 25 2023 - 06:48:57 EST


On Thu, 11 May 2023, Jerome Neanne wrote:

> Use new API for power-off mode support:
> Link: https://lwn.net/Articles/894511/
> Link: https://lore.kernel.org/all/7hfseqa7l0.fsf@xxxxxxxxxxxx/
>
> sys-off API allows support of shutdown handler and restart handler.
>
> Shutdown was not supported before that enhancement.
> This is required for platform that are not using PSCI.
>
> Test:
> - restart:
> # reboot
> Default is cold reset:
> # cat /sys/kernel/reboot/mode
> Switch boot mode to warm reset:
> # echo warm > /sys/kernel/reboot/mode
> - power-off:
> # halt
>
> Tested on AM62-LP-SK board.
>
> Signed-off-by: Jerome Neanne <jneanne@xxxxxxxxxxxx>
> Suggested-by: Andrew Davis <afd@xxxxxx>
> Reviewed-by: Andrew Davis <afd@xxxxxx>
> ---
>
> Notes:
> Change-log v2 to v1
> v1: Link: https://lore.kernel.org/all/20230203140150.13071-1-jneanne@xxxxxxxxxxxx/
> Andrew Davis Review:
> - Use new helpers devm_register_restart_handler and devm_register_power_off_handler
> Vignesh Raghavendra:
> - Fix typo on board name in commit message
>
> drivers/mfd/tps65219.c | 41 ++++++++++++++++++++++++++++++-----------
> 1 file changed, 30 insertions(+), 11 deletions(-)

Couple of nits.

> diff --git a/drivers/mfd/tps65219.c b/drivers/mfd/tps65219.c
> index 0e402fda206b..5115d0a66701 100644
> --- a/drivers/mfd/tps65219.c
> +++ b/drivers/mfd/tps65219.c
> @@ -25,25 +25,34 @@ static int tps65219_cold_reset(struct tps65219 *tps)
> TPS65219_MFP_COLD_RESET_I2C_CTRL_MASK);
> }
>
> -static int tps65219_restart(struct notifier_block *this,
> - unsigned long reboot_mode, void *cmd)
> +static int tps65219_soft_shutdown(struct tps65219 *tps)
> {
> - struct tps65219 *tps;
> + return regmap_update_bits(tps->regmap, TPS65219_REG_MFP_CTRL,
> + TPS65219_MFP_I2C_OFF_REQ_MASK,
> + TPS65219_MFP_I2C_OFF_REQ_MASK);
> +}
>
> - tps = container_of(this, struct tps65219, nb);
> +static int tps65219_power_off_handler(struct sys_off_data *data)
> +{
> + tps65219_soft_shutdown(data->cb_data);
> + return NOTIFY_DONE;
> +}
>
> +static int tps65219_restart(struct tps65219 *tps,
> + unsigned long reboot_mode)

Why the line-wrap?

> +{
> if (reboot_mode == REBOOT_WARM)
> tps65219_warm_reset(tps);
> else
> tps65219_cold_reset(tps);
> -

This has nothing to do with the patch, and I liked it better before.

> return NOTIFY_DONE;
> }
>
> -static struct notifier_block pmic_rst_restart_nb = {
> - .notifier_call = tps65219_restart,
> - .priority = 200,
> -};
> +static int tps65219_restart_handler(struct sys_off_data *data)
> +{
> + tps65219_restart(data->cb_data, data->mode);
> + return NOTIFY_DONE;
> +}
>
> static const struct resource tps65219_pwrbutton_resources[] = {
> DEFINE_RES_IRQ_NAMED(TPS65219_INT_PB_FALLING_EDGE_DETECT, "falling"),
> @@ -269,13 +278,23 @@ static int tps65219_probe(struct i2c_client *client)
> }
> }
>
> - tps->nb = pmic_rst_restart_nb;
> - ret = register_restart_handler(&tps->nb);
> + ret = devm_register_restart_handler(tps->dev,
> + tps65219_restart_handler,
> + tps);
> +
> if (ret) {
> dev_err(tps->dev, "cannot register restart handler, %d\n", ret);
> return ret;
> }
>
> + ret = devm_register_power_off_handler(tps->dev,
> + tps65219_power_off_handler,
> + tps);
> + if (ret) {
> + dev_err(tps->dev, "failed to register power-off handler: %d\n",
> + ret);

I wouldn't wrap here either to be honest.

checkpatch.pl now complains at 100-chars.

> + return ret;
> + }
> return 0;
> }
>
> --
> 2.34.1
>

--
Lee Jones [李琼斯]