Hello,
In the past we have utilised the system bootloader on the STM32F429ZIT6, to implement a command to allow firmware upgrades to be completed by the end user. As part of this, the command would configure an LED Driver IC to update an RGB status LED, giving the user an indication that they are in bootloader mode.
This works great, and led to the assumption that the system bootloader does not reconfigure pins that are not used as by any of the interfaces listed in the application note AN2606, since if this was the case, the LED would not stay on since an enable pin needs to be held high by the STM32.
We are now looking to implement the same firmware upgrade command on a different product that uses the STM32G473VET6TR instead, and have not had as much success.
It appears as though once you enter the system bootloader, the LED enable pin (PD2) is reconfigured since the line is no longer being driver high by the STM32.
For some context, I have included the function that we are using to switch to the system bootloader from our application code below, and please note that before this function is called the LED is configured and confirmed to be on. However once I step into SysMemBootJump(), the LED enable line is no longer driven high and hence the LED is turned off.
Also just to confirm once this function is executed, the product does enter bootloader mode and can be connected to and programmed by the STM32Programmer tool.
Seeing as I have been unable to find any documentation stating what the System Bootloader does with unused pins, I wanted to reach out for some guidance.
Can we rely on the system bootloader to not reconfigure pins not listed in AN2606?
Many thanks
Calum
static void EnterFactoryBootloader(void)
{
//Set the stack and privilege levels - thread mode, default stack pointer MSP used.
__set_CONTROL(0);
// STM Doc AN2606 specifies the system memory location for the STM32F429 as 0x1FFF0000 (Same applies to STM32G47xx)
// When Cortex-M boots it expects to find the vector table at this address.
// The first thing in the vector table is the address of its stack
// and the second the address of the program entry point (offset 4 bytes).
volatile const uint32_t sysMemAddress = 0x1FFF0000;
volatile const uint32_t progEntryPoint = sysMemAddress + 4;
volatile const uint32_t mspAddress = (*((uint32_t *)sysMemAddress)); //=0x20002d40
//delclare a void function pointer
void (*SysMemBootJump)(void);
//set the pointer to the jump location
SysMemBootJump = (void (*)(void)) (*((uint32_t *) progEntryPoint));
//Disable the SysTick timer and reset it to default values
SysTick->CTRL = 0; //reset systic timer
SysTick->LOAD = 0;
SysTick->VAL = 0;
//set vector table offset to zero
SCB->VTOR = 0;
//Remap system memory to address 0x0000-0000
__HAL_SYSCFG_REMAPMEMORY_SYSTEMFLASH();
//Set the main stack pointer to its default values
//Note this must be done last as the local variables of this function use the stack
//__set_MSP(0x20002d40);
__set_MSP(mspAddress);
//do the jump
SysMemBootJump(); // <-- ** LED stays on up to here! **
//code should never get here
while(1);
}
