stm32L051 registers don't change

[EPora.1]

Hello!

I'm working on a custom board with an stm32L051 and the HSI16 crystal. I'm setting the clock registers, Flash, PWR and then I try to enable the clock for TIM21. as you can see in the code I set the bit but the computer "doesn't listen"(I am looking in the APB2ENR reg in the debugging env). anyone had this issue?

adding the clock diagram and code.

Thanks,

Eyal Porat

The clock setup and FLASH+PWR

void clockSetup(void)
{
	
	RCC->CR |= (1 << 0); //set the HSI16ON
	while(!(RCC->CR & (1 << 2))); //wait for the HSI16RDYF to set
	
	RCC->APB1ENR |= (1 << 28); //set the POWERN bit
	
	PWR->CR |= (1 << 11); //set VOS bits to 0x1
	
	FLASH->ACR |= (1 << 0); //set the LATENCY bit to 1WS(2cpu cycles)
	
	FLASH->ACR |= (1 << 1); //set the PERFTEN bit
	
	FLASH->ACR |= (1 << 6); //set the PRE_READ bit
	
	uint32_t temp = FLASH ->ACR; //it is recommended to read the register
	
	
	RCC->CFGR = 0;
	
	RCC->CFGR &= ~(15 << 4); //set the AHB prescaler to 1
	
	RCC->CFGR &= ~(7 << 8); //set the PPRE1/APB1/PCLK1 to div 1
	
	RCC->CFGR &= ~(7 << 11); //set the PPRE2/APB2/PCLK2 to div 1
	
	RCC->CFGR |= (1 << 18); //set the PLLMUL scaler to 4
	
	RCC->CFGR |= (1 << 22); //set the PLLDIV devidor to 2
	
	
	RCC->CFGR &= ~(1 << 16); //set the input for the PLL to HSI16
	
	RCC->CR |= (1 << 24); //set the PLLON bit to enable the PLL
	
	while(!(RCC->CR & (1 << 25))); //wait for the PLL to get ready
	
	
	RCC->CFGR |= (3 << 0); //set the SW bits to 01 (HSI16 input)
	
	while(!(RCC->CFGR & (3 << 2))); //wait for the clock source to update
	
	//SystemCoreClockUpdate();
	
}

The Timer functions

void TIM21conf(void)
{
	//RCC->APB2ENR |= (1 << 2); //enable the TIM21 clock
	RCC->APB2ENR |= RCC_APB2ENR_TIM21EN;
	while(TIM21->ARR == 0);
	//timers clock is 32mhz
	TIM21->PSC = 32000-1; //clock fraquency -1 to get a 1khz pulse
	TIM21->ARR = 0xFFFF; //set the maximum value to be the max possible
	
	TIM21->CR1 |= (1 << 0); //enable the timer
	
	while (!(TIM21->SR & (1 << 0))); //wait for the timer ready bit to set 
}
 
void delayMs(uint32_t millis)
{
	TIM21->CNT = 0; //reset the count value
	
	while (TIM21->CNT <= (millis)); //wait for the timer to reack the pre-set value
}

int main()
{
	clockSetup();
	
	SystemCoreClockUpdate();
	
	TIM21conf();
 
	while(1)
	{
		
	}
}

the main code.

[waclawek.jan]

Where is the program counter? Does it reach instructions which set APB2ENR at all?

JW

[EPora.1]

Hi thanks for the answer!

What does the instruction looks like? can it be dew to optimization? I'm using -O1

[waclawek.jan]

I mean, check, if your program gets through the clock setup. For example, place a breakpoint at the line where APB2ENR is written, does the program get there?

Generally, single-step the program and observe registers which are written.

Also note, that PWR_CR.is nonzero after reset. Reading it out and checking against RM would indicate the potential problem.

JW

[EPora.1]

Hi, I have tested it in a step by step way and sometimes it reaches the line and sets the reg, sometimes it reaches and doesn't set and sometimes it glitches to 0xFFFFFFFE. also PWR_CR is nonzero, the VOS bits are set by default,

Eyal

[Tesla DeLorean]

The optimizer shouldn't fold _IO (volatile) memory accesses.

One hazard, especially with slower peripherals, is ensuring that the write buffers clear, and the clock enable latches in the RCC *before* you touch any synchronous logic dependent on the clocks running.

[EPora.1]

How can I check that?

Thanks, Eyal