How to store code into two discontinuous flash area automatically in CubeIDE?

[DZhih.1]

STM32G474RC，flash size is 256kB，default dual bank，so the flash size is 0x08000000~0x08020000, 0x080400000~0x08060000.

If the code size is larger than 128kB, but the address is non-contiguous.

How to store the exceeded programs into Bank2 automatically?

Is there any method to configure the link file in CubeIDE?

I know i can specify some function to certain address, but it's inconvenient at all.

[TDK]

The GCC linker isn't set up to do this, unfortunately.

It was a bizarre choice to have a discontinuous flash.

[MM..1]

I mean you can in linker script define .o files for specific areas, or you need add this to source code.

[alister]

Is 0x08000000 or 0x08040000 mirrored at 0x08020000?

Otherwise declare two memory regions per https://sourceware.org/binutils/docs/ld/MEMORY.html and manually place your .text and/or .rodata in them. There are several ways. Per function atribute like __attribute__((section(".myOtherbank")) void MyFunction(void) {}, individually listing the objects of the input sections of the linker command file, etc. It's all in the docs. https://gcc.gnu.org/onlinedocs/gcc-4.7.2/gcc/, https://sourceware.org/binutils/docs/ld/index.html.

[DZhih.1]

It seems that manually place my function in this two memory is the only solution now. IAR and TI CCS both support this operation. CubeIDE can not support it.

[DZhih.1]

Yes, the dual bank is discontinuous unless single bank is set.

[alister]

It's using GCC. Read the links I posted. CubeIDE has nothing to do with it.