STM32 timer delay - assembler

Dear ARM-Community,

I need help on my timer implementation. I want to let the LED flash with 1Hz on my STM32F0-Nucleo with an 48MHz internal clock.

My C code is working perfectly:

...

#define LD3 (0x03)

int main() {

  //setup LED
  RCC->AHBENR |= RCC_AHBENR_GPIOBEN;
  GPIOB->MODER  &= ~(0x3 << (LD3*2));
  GPIOB->MODER  |=  (0x1 << (LD3*2));
  GPIOB->OTYPER &= ~(1 << LD3); 

  RCC->APB1ENR |= RCC_APB1ENR_TIM2EN;
  TIM2->CR1 &= ~(TIM_CR1_CEN);

  TIM2->PSC = 47999;
  TIM2->ARR = 499;
  
  TIM2->CR1 |= TIM_CR1_CEN;
  TIM2->EGR |= TIM_EGR_UG;

  while(1) {
    if (TIM2->SR & TIM_SR_UIF) {
        GPIOB->ODR ^= (1UL << LD3);
        TIM2->SR &= ~(TIM_SR_UIF);
      }
  }
}

Here is my assembler code, which does not work 100%:

...

@enable TIM2
	ldr r0, =APB1ENR
	movs r1, #0x1
	str r1, [r0]
@reset TIM2 clk
	ldr r0, =TIM2_CR1
	orrs r1, r1 
	str r1, [r0]
@set output cmp mode
	ldr r0, =TIM2_CCMR1
	movs r1, #0x7
	lsls r1, r1, #0x4
	str r1, [r0]
@set prescaler	
	ldr r0, =TIM2_PSC
	eors r1, r1
	ldr r1, =#47999
	@ldr r1, =#0xbb7f
	str r1, [r0]
@set auto reload value
	ldr r0, =TIM2_ARR
	eors r1, r1
	ldr r1, =#1000			@500ms - 1ms = 0x1f3
	str r1, [r0]
@set TIM2 clk
	ldr r0, =TIM2_CR1
	movs r1, #0x1
	str r1, [r0]

@setup vectors
	ldr r0, =TIM2_SR
	ldr r1, =GPIOB_ODR
	movs r2, #0x1
	movs r3, #0x8
@trigger update event
	ldr r6, =TIM2_EGR
	movs r7, #0x1
	str r7, [r0]

@while loop
w0:	
	ldr r4, [r0]	
	tst r4, r2
	beq w0

	ldr r5, [r1]
	eors r5, r5, r3
	str r5, [r1]
@reset SR
	eors r5, r5
	str r5, [r0]
	b w0

I tested the LED already with a simple loop, that worked. All the vectors are also set correctly! But now, even though I set the right values, the delay is not working correctly.

I guess it has strangely to do something with the ARR, because I can change the f with it.

I have no clue what is wrong here. Can somebody help me?

Thank you!