STM32H563: BANK SWAP

[jowakar122]

I am using NUCLEO-H563ZI and am trying to implement bank swap for firmware update. I am using YMODEM to update firmware. Initially when my code is being executed from BANK 1 and after writing the new firmware at BANK 2. The code execution starts from BANK 2 successfully. But the reverse in not happening. After all the process I checked the address where I am writing the new firmware in BANK 1 and nothing was there which cause hard fault. 

Therefore, why the process doesn't fail initially when writing firmware at BANK 2 and it fails only when writing at BANK 1 even though the code is same. Below I have attached some files. 

void SerialDownload(void)
{
 uint8_t number[11] = {0};
 uint32_t size = 0;
 COM_StatusTypeDef result;

 Serial_PutString((uint8_t *)"\rWaiting for the file to be sent ... (press 'a' to abort)\n\r");
 result = Ymodem_Receive( &size );
 if (result == COM_OK)
 {
	 if ((erase_operation==1) && (end_of_download_operation ==0))
	 {
		 end_of_download_operation = 1;
	 }
	 Serial_PutString((uint8_t *)"\n\n\r Programming Completed Successfully!\n\r-----------------------\r\n Name: ");
	 Serial_PutString(aFileName);
	 Int2Str(number, size);
	 Serial_PutString((uint8_t *)"\n\r Size: ");
	 Serial_PutString(number);
	 Serial_PutString((uint8_t *)" Bytes\r\n");
	 Serial_PutString((uint8_t *)"-----------------------\n");
 }
 else if (result == COM_LIMIT)
 {
	 Serial_PutString((uint8_t *)"\n\n\rThe image size is higher than the allowed space memory!\n\r");
 }
 else if (result == COM_DATA)
 {
	 Serial_PutString((uint8_t *)"\n\n\rVerification failed! Error while writing to the Flash\n\r");
 }
 else if (result == COM_ABORT)
 {
	 Serial_PutString((uint8_t *)"\r\n\nAborted by user.\n\r");
 }
 else if (result == ERASE_ERR)
 {
	 Serial_PutString((uint8_t *)"\r\n\nERROR! Erasing the Flash.\n\r");
 }
 else
 {
	 Serial_PutString((uint8_t *)"\n\rFailed to receive the file!\n\r");
 }
}

 

[jowakar122]

Does it have to do anything with ICACHE?

 

[jowakar122]

Ok I got it. In my case Bank 1 is at 0x08000000 and Bank 2 is at 0x08100000.  So when the SWAP_BANK bit is set it remaps the logical addresses. After reset code is being executed from Bank 2, it is remapped at 0x08000000 and Bank 1 is at 0x08100000. So I was writing the new firmware at 0x08000000 from where the code is being generated so it would fail.

Therefore, whenever I get new firmware I have to write it at 0x08100000. For more reference go through this article:
STM32H5_USB_DFU_Dual_Bank_demo/STM32H5 USB DFU Dual Bank.pdf at main · stm32-hotspot/STM32H5_USB_DFU_Dual_Bank_demo