EMAC driver initializes an incorrect pin (PG2) on the STM32F4 -> NUCLEO_F429ZI target

[akiroz]

Description of defect

EMAC driver initializes an incorrect pin (PG2) on the STM32F4 -> NUCLEO_F429ZI target

In mbed 5.15, the EMAC driver initializes PG2 for RMII RX ER on STM32F429, this pin is not part of the RMII interface for this chip.

See:

mbed-os/features/netsocket/emac-drivers/TARGET_STM/TARGET_STM32F4/TARGET_NUCLEO_F429ZI/stm32f4_eth_init.c

Line 54 in bacf378

RMII_MII_RXER ---------------------> PG2

The correct pin for the RMII RX ER signal is on PB10, however this signal is not connected on the Nucleo development board.

Target(s) affected by this defect ?

STM32F4 -> NUCLEO_F429ZI

I suspect this issue also affects many other STM32 targets but I have yet to verify this.

Toolchain(s) (name and version) displaying this defect ?

N/A

What version of Mbed-os are you using (tag or sha) ?

mbed-os-5.15.2

What version(s) of tools are you using. List all that apply (E.g. mbed-cli)

N/A

How is this defect reproduced ?

N/A

[jeromecoutant]

Hi
I checked User Manual for this Nucleo, and there is no RMII RX ER (not on PG2, not in PB10)...

[akiroz]

Yes, the RMII RX ER signal is not routed in the Nucleo board. PB10 is where the MCU would support this signal if used.

[akiroz]

Since this signal is optional and not routed in Nucleo, I think an easy fix is simply deleting PG2 from the init code.
I checked a few other STM targets and they also seem to be affected by this issue.

[jeromecoutant]

Thx, please propose a patch in order to review and correct :-)