Re: [RFC PATCH 2/2] leds: flash: Add support for the TI LM3643 dual LED flash driver

[Lee Jones <lee@kernel.org>]

On Fri, 21 Aug 2026, sashiko-bot@kernel.org wrote:

> Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
> - [High] Turning off an LED fails to clear the chip mode bits, causing power drain and breaking flash strobe teardown.

Please review, explain or fix the Sashiko review(s).

If this is a set, this request includes the other patches too.

> --
> 
> --- Patch [2]: [RFC PATCH 2/2] leds: flash: Add support for the TI LM3643 dual LED flash driver ---
> commit bdf1945ab184ca005c2f401ecd5408571ece1de3
> Author: Rillian Grant <rillian.grant@gmail.com>
> 
> leds: flash: Add support for the TI LM3643 dual LED flash driver
>     
> The LM3643 is an I2C dual LED flash driver. Its two high-side current
> sources share a synchronous boost converter and each supports torch and
> flash modes with programmable current and flash timeout alongside fault
> reporting.
> 
> > diff --git a/drivers/leds/flash/leds-lm3643.c b/drivers/leds/flash/leds-lm3643.c
> > new file mode 100644
> > --- /dev/null
> > +++ b/drivers/leds/flash/leds-lm3643.c
> 
> [ ... ]
> 
> > +static int lm3643_torch_set_brightness(struct led_classdev *led_cdev,
> > +				       enum led_brightness brightness)
> > +{
> 
> [ ... ]
> 
> > +	if (brightness == 0)
> > +		return regmap_update_bits(led->regmap, LM3643_REG_ENABLE,
> > +				chan->enable_bit, 0);
> 
> [Severity: High]
> When brightness is 0, this only clears the channel's enable bit but leaves
> the LM3643_MODE_MASK bits intact. If no other LEDs are active, does this
> leave the chip in an active power state (Torch or Flash mode) instead of
> transitioning to LM3643_MODE_STANDBY, potentially causing continuous battery
> drain?
> 
> [ ... ]
> 
> > +static int lm3643_flash_strobe_set(struct led_classdev_flash *fled_cdev, bool state)
> > +{
> 
> [ ... ]
> 
> > +	if (!state) {
> > +		if (mode != LM3643_MODE_FLASH || !(enable_reg & chan->enable_bit))
> > +			return 0;
> > +
> > +		return regmap_update_bits(led->regmap, LM3643_REG_ENABLE,
> > +			LM3643_MODE_MASK, LM3643_MODE_STANDBY);
> > +	}
> 
> [Severity: High]
> If the torch brightness was set to 0 during a flash, the channel's enable
> bit will have been cleared. During the subsequent teardown call to
> lm3643_flash_strobe_set() with state = false, does this early return trigger
> because the enable bit is no longer set?
> 
> This would fail to restore the mode to LM3643_MODE_STANDBY, permanently
> stranding the driver's state machine in Flash mode.
> 
> -- 
> Sashiko AI review · https://sashiko.dev/#/patchset/20260821083620.68324-1-rillian.grant@gmail.com?part=2

-- 
Lee Jones