[STM32F405] flash strange behaviour
Hello all,
I'm trying to lock the flash in level 1 mode. To do so, I'm using the provided ST APIs. In particular, I'm doing:
FLASH_Unlock(); // unlock flash to deal with it
FLASH_OB_Unlock(); // unlock option byte to deal with it
if(FLASH_OB_GetRDP() != SET) // check current protection status
{
FLASH_OB_RDPConfig(OB_RDP_Level_1);
FLASH_OB_Launch(); // launch the option byte writing procedure
}
FLASH_OB_Lock(); // lock option byte
FLASH_Lock(); // lock flash
Now, looking at the function implementations of the library, I can see something I'm not understanding in FLASH_GetStatus:
FLASH_Status FLASH_GetStatus(void)
{
FLASH_Status flashstatus = FLASH_COMPLETE;
if((FLASH->SR & FLASH_FLAG_BSY) == FLASH_FLAG_BSY)
{
flashstatus = FLASH_BUSY;
}
else
{
if((FLASH->SR & FLASH_FLAG_WRPERR) != (uint32_t)0x00)
{
flashstatus = FLASH_ERROR_WRP;
}
else
{
if((FLASH->SR & (uint32_t)0xEF) != (uint32_t)0x00)
{
flashstatus = FLASH_ERROR_PROGRAM;
}
else
{
if((FLASH->SR & FLASH_FLAG_OPERR) != (uint32_t)0x00)
{
flashstatus = FLASH_ERROR_OPERATION;
}
else
{
flashstatus = FLASH_COMPLETE;
}
}
}
}
/* Return the FLASH Status */
return flashstatus;
}
Specifically, it is checking the SR register with mask EF. Looking at datasheet it means that we are taking into consideration also the reserved bits an the bit 0.
I created an other project using stm32cubeIDE version 1.15.0 and found out that in the new version of library the implementation is a bit different
if(__HAL_FLASH_GET_FLAG((FLASH_FLAG_OPERR | FLASH_FLAG_WRPERR | FLASH_FLAG_PGAERR | \
FLASH_FLAG_PGPERR | FLASH_FLAG_PGSERR)) != RESET)
The library does not check reserved bits and does not check EOF bit.
So, what bits should I check?
With debugger I read the content of SR register and the value is 0xC0 which means that I have the PGSERR and PGPERR bits set to 1.
Now, If I ignore that 2 bits everything work as expected and the flash get protected but I need to understand If I can safely ignore them and why that bits are set? The procedure I'm performing should be correct.
Is there something documented in some ERRATA?
Thanks for the answer.
Marco.