boards: disco_l475_iot1: fix arduino_i2c assignment

[glenn-andrews]

0f05f58 assigned the arduino_i2c to the incorrect pins for the Arduino UNO connector. This assigns them back.

Fixes #88558

[kartben]

Sorry that I broke this, and probably better to revert at this point.

FWIW I was actually using a shield (https://www.mikroe.com/arduino-uno-click-shield) with effectivley a Uno R1 form factor (it doesn't even have D18/D19) and for which I2C is expected to be on A4/A5.
It almost feels as if we need to add a dt binding for Arduino Uno R1 headers in addition to R3 that we have right now to avoid these kind of issues in the future, and possibly that shield I MikroE shield should not "pretend" it's Uno R3 compatible (or the issue is STM32 specific and it should not have D18 D19 expose something different than A4/A5 if the expectation for R3 is that they're effectively the same)

@erwango might be able to comment or pull in folks that know

[glenn-andrews]

I think you're right in adding an Arduino UNO R1 binding as well as the existing R3 binding

[glenn-andrews]

And this looks like the updated R3 click shield ;)

https://www.mikroe.com/click-shield-for-arduino-uno

[mathieuchopstm]

Sorry that I broke this, and probably better to revert at this point.

FWIW I was actually using a shield (https://www.mikroe.com/arduino-uno-click-shield) with effectivley a Uno R1 form factor (it doesn't even have D18/D19) and for which I2C is expected to be on A4/A5. It almost feels as if we need to add a dt binding for Arduino Uno R1 headers in addition to R3 that we have right now to avoid these kind of issues in the future, and possibly that shield I MikroE shield should not "pretend" it's Uno R3 compatible (or the issue is STM32 specific and it should not have D18 D19 expose something different than A4/A5 if the expectation for R3 is that they're effectively the same)

I would argue the answer to the question is on this very diagram, which only lists I²C as available on pins D18/D19. On the other hand, the Arduino Uno R3 datasheet does specify that D18/D19 are duplicated A4/A5, at page 11.

I don't know if there are official guidelines from Arduino about which pins shields should be used for what purpose, but I would remark that all official Arduino shields listed as compatible with Uno R3 use D18/D19 for this purpose - see schematics.

To answer your second question, I would argue that the question should really be "why are D18/D19 shorted to A4/A5 on Arduino R3?", the answer to which becomes obvious when looking at the ATmega328p's pinout 🙂

(To answer more seriously: as far as I know (i.e., this is not an official statement from ST), D18/D19 and A4/A5 are routed to different pins (when possible) because this allows using more peripherals at the same time)

TL;DR: in my opinion, this is a shield issue for using A4/A5 rather than D18/D19.