Re: [PATCH 4/4] regulator: fan53555: Add support for mode operations on Silergy devices

[sashiko-bot@kernel.org]

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:

Pre-existing issues:
- [High] `fan53555_set_mode()` writes to `di->vol_reg` instead of `di->mode_reg` when setting `REGULATOR_MODE_NORMAL`, causing severe state corruption for several vendors.
--

--- Patch [4]: [PATCH 4/4] regulator: fan53555: Add support for mode operations on Silergy devices ---
commit f660e0b49b4dfc8b2006ec4509cf598aba57babb
Author: Victor Krawiec <victor.krawiec@arturia.com>

regulator: fan53555: Add support for mode operations on Silergy devices

This patch maps Silergy PMIC device modes via device tree by mapping
SYR82X_REGULATOR_MODE_AUTO to REGULATOR_MODE_NORMAL. It exposes configurable
PWM modes for these regulators.

> diff --git a/drivers/regulator/fan53555.c b/drivers/regulator/fan53555.c
> index c282236959b18..8664659f0d809 100644
> --- a/drivers/regulator/fan53555.c
> +++ b/drivers/regulator/fan53555.c

[ ... ]

> @@ -389,6 +390,18 @@ static int rk8602_voltages_setup_rockchip(struct fan53555_device_info *di)
>  	return 0;
>  }
>  
> +static inline unsigned int fan53555_silergy_map_mode(unsigned int mode)
> +{
> +	switch (mode) {
> +	case SYR82X_REGULATOR_MODE_FORCE_PWM:
> +		return REGULATOR_MODE_FAST;
> +	case SYR82X_REGULATOR_MODE_AUTO:
> +		return REGULATOR_MODE_NORMAL;

[Severity: High]
This isn't a bug introduced by this patch, but returning REGULATOR_MODE_NORMAL
here exposes a pre-existing issue in fan53555_set_mode() that affects multiple
vendors (such as FAN53526, TCS, or RK8602).

When an in-kernel PM consumer driver or sysfs interface calls
regulator_set_mode() to set REGULATOR_MODE_NORMAL, it incorrectly updates
di->vol_reg instead of di->mode_reg:

drivers/regulator/fan53555.c:fan53555_set_mode() {
	...
	case REGULATOR_MODE_NORMAL:
		regmap_update_bits(rdev->regmap, di->vol_reg, di->mode_mask, 0);
		break;
	...
}

For RK8602, di->mode_mask is VSEL_MODE (BIT 6), but in vol_reg (e.g.,
RK8602_VSEL1), bit 6 represents a 64-step voltage increment (400mV). Clearing
this bit inadvertently plummets the physical voltage, which can cause hardware
brownouts, instant kernel panics, or memory corruption.

Is it possible to fix this existing issue in fan53555_set_mode() by using
di->mode_reg instead of di->vol_reg before or alongside this change?

> +	default:
> +		return REGULATOR_MODE_INVALID;
> +	}
> +}

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260715155141.128601-1-victor.krawiec@arturia.com?part=4