[PATCH linux dev-5.10 06/35] ARM: dts: aspeed: rainier: Add leds that are off PCA9552

[Joel Stanley]

On Fri, 12 Mar 2021 at 00:21, Milton Miller II <miltonm at us.ibm.com> wrote:
>
>
>
> -----"openbmc" <openbmc-bounces+miltonm=us.ibm.com at lists.ozlabs.org> wrote: -----
>
> >To: Eddie James <eajames at linux.ibm.com>
> >From: Joel Stanley
> >Sent by: "openbmc"
> >Date: 03/11/2021 06:09PM
> >Cc: OpenBMC Maillist <openbmc at lists.ozlabs.org>
> >Subject: [EXTERNAL] Re: [PATCH linux dev-5.10 06/35] ARM: dts:
> >aspeed: rainier: Add leds that are off PCA9552
> >
> >On Mon, 8 Mar 2021 at 22:54, Eddie James <eajames at linux.ibm.com>
> >wrote:
> >>
> >> From: Vishwanatha Subbanna <vishwa at linux.vnet.ibm.com>
> >>
> >> These LEDs are on the fans and are connected via a
> >> pca9551 i2c expander
> >
> >This change doesn't make sense. The pca9551 is an i2c LED expander,
> >so
> >we don't need to expose the pins as GPIOs and then attach a gpio-leds
> >driver to them. We should instead simply configure the pca955x driver
> >to drive the LEDs as LEDs.
>
> I'll refresh your memory on why we have been doing this in our
> devie trees and then let you consider if this is desired or not.
>
> The led system insistes on creating a compact map (no holes) (as
> does the reset subsystem).
>
> However, this means the relative led number for a pin changes
> as the prior pins change from gpio to led configuration.
>
> For example if pins 2 and 7 are leds, they become leds 0 and 1.
> Changing pin 5 to also be an led means that pin 7 is now led 2
> not led 1 on the led subsystem.

Thanks for the rationale reminder.

Are these led numbers important to userspace, or does the renumbering
affect device tree changes only?