STM32L476RE based Disco board not running Micropython

[kowshik1234]

Hi,

I have compiled the micropython from the source by making changes to the UART port on which REPL should launch and it generated a .hex file.

I want to flash this to the MCU but I want to do it with STLink/V2 debugger. I tried flashing it but I don't see the REPL running on this UART port. What is the recommended way to flash the micropython hex files onto STM32L476RE MCU?

FYI, I have this Eval board - ST25R3911B-DISCO - Discovery kit for the ST25R3911B high performance HF reader / NFC initiator - STMicroelectronics

I am trying to do a porting of the NFC libraries of this board to micropython and as a part of my contribution task I tried doing the above. Please help me how to tweak and run the micropython.

Thanks,
G Kowshik

[davefes]

In my attempts to get REPL working on a STM32WLE5 I was told that REPL uses the LPUART1_TX and LPUART1_RX pins. I then use a ST-Link V2, STM32 Cube Programmer and the SWCLK and SWDIO pins to program it.

[kowshik1234]

@robert-hh

Which loader script do you use?

I have enabled verbose build and from there I see the script is using the stm32l476xg.ld and it has the flash sections named as below

/* Specify the memory areas */
MEMORY
{
    FLASH (rx)      : ORIGIN = 0x08000000, LENGTH = 512K /* sectors 0-255 */
    FLASH_FS (r)    : ORIGIN = 0x08080000, LENGTH = 512K /* sectors 256-511 */
    RAM (xrw)       : ORIGIN = 0x20000000, LENGTH = 96K
    SRAM2 (xrw)     : ORIGIN = 0x10000000, LENGTH = 32K /* not contiguous with RAM */
}

_micropy_hw_internal_flash_storage_start = ORIGIN(FLASH_FS);
_micropy_hw_internal_flash_storage_end   = ORIGIN(FLASH_FS) + LENGTH(FLASH_FS);

[robert-hh]

It should use the stm32l476xe.ld script, matching the STM32L476RE device.

[kowshik1234]

@robert-hh I have taken a reference example from the STM's SDK and this is what I found from their linker script, seems the current one that is being picked by Micropython makesystem is correct only. What do you think please?

Regardless, I will give a try by plugging in the stm32l476xe.ld values to the existing script and see how it goes. I remember trying this out and the execution gave a hard fault in the mp_init( ) //Initialize micropython function, however let me try again.

Also on a lighter note, I'm newly getting this warning when build, Is it complaining about the linker file changes that has happened and the newly build binary has changes? I've no idea about this.

GEN build-NUCLEO_L476RG/frozen_content.c
CC build-NUCLEO_L476RG/frozen_content.c
LINK build-NUCLEO_L476RG/firmware.elf
   text    data     bss     dec     hex filename
 405828      96   27588  433512   69d68 build-NUCLEO_L476RG/firmware.elf
GEN build-NUCLEO_L476RG/firmware.bin
GEN build-NUCLEO_L476RG/firmware.hex
GEN build-NUCLEO_L476RG/firmware.dfu
make: warning:  Clock skew detected.  Your build may be incomplete.

Thanks

[robert-hh]

I do not know what the warning means. Sometimes it is the best to clean the build directory and make a complete new build.

As of the loader script defines: They tell, how much of the flash is used for code, and where the file system starts. The hard fault you have seen in debugging happened in accessing the file system in the FLASH_FS area, which according to your loader settings was outside of the existing flash. The stm32l476xe.ld settings move it into the existing flash range of 512k.

[kowshik1234]

@robert-hh Thanks alot for the help. I am able to boot up the REPL after changing the linker script to xe.ld, really appreciate your help.

Thanks again, Cheers !!