Re: [PATCH] leds: flash: Add driver for the TI LM3554

From: Andy Shevchenko

Date: Wed Nov 18 2026 - 02:33:12 EST


On Tue, Nov 17, 2026 at 05:57:45PM +0000, Volodymyr Chapman wrote:
> Hi Lee, Pavel,
>
> Andy Shevchenko pointed out, in review of a related camera series, that
> drivers/leds/leds-lm355x.c already lists the LM3554. I should have
> mentioned it in the commit message.
>
> It can't be used on the Lenovo ThinkPad 10 as it is: it only probes with
> platform data (no ACPI or DT matching, and nothing in the tree provides
> the platform data), it does not drive the HWEN enable line, which has to
> be raised here before the chip responds, and it registers three plain
> LED class devices rather than a flash LED class device, so there is no
> flash timeout or fault reporting.
>
> This patch follows commit f1aa35ab5943 ("media: atomisp: Drop the
> atomisp custom lm3554 flash driver"), which removed the atomisp-specific
> LM3554 driver for the same hardware and said that it should be
> re-implemented with the led-class-flash API if needed.
>
> Would you prefer a separate flash class driver for the LM3554 as in this
> patch, or extending leds-lm355x (fwnode/ACPI matching, the enable GPIO
> and the flash class)? I'll do whichever you prefer in v2, together with
> the fixes for the issues the Sashiko review found.

Since lm355x has no fwnode support, I think we have no users for it
in kernel (there are no users of platform data either).

I would move the driver to the flash subfolder (and rename it back to lm3556
as all those 'x' variants showed quite a confusion for the users and developers
as we noticed in IIO and have now even requirement to not make any driver names
with wildcard in it) and develop necessary support.

Also the documentation Documentation/leds/leds-lm3556.rst needs to be amended.

--
With Best Regards,
Andy Shevchenko