.syntax unified         @ Use the unified syntax

.thumb                  @ Use the Thumb instruction set

.global leds_blink
.global delay_loop
.global leds_init
.global main

.equ	rcc_reg,            0x40023800   // adresa registru RCC
.equ	rcc_ahb1enr_offset, 0x30         // offset registru AHB1ENR
.equ	rcc_gpiob_en,       0x0002       // bit pro povoleni hodin pro GPIOB
.equ	rcc_gpioa_en,			  0x0001       // bit pro povoleni hodin pro GPIOB

.equ 	gpio_a,             0x40020000   // adresa registru GPIOA
.equ 	gpio_b,             0x40020400   // adresa registru GPIOA
// .equ 	gpio_c,             0x40020800   // adresa registru GPIOB

.equ	gpio_moder_offset,  0x00         // offset adresy registru MODER
.equ	gpio_otyper_offset, 0x04         // offset adresy registru OTYPER
.equ	gpio_odr_offset,	  0x14         // offset adresy registru ODR

.equ	gpioa_odr_init,     (0 << 5)   // inicializacni hodnota pro registr ODR (turn off upper 4 leds)

// Initialization values for moder, odr, otyper
.equ	gpioa_moder_leds,   (0x01 << (2 * 5)) | (0x01 << (2 * 6)) | (0x01 << (2 * 7))
.equ	gpiob_moder_leds,   (0x01 << (2 * 6))
.equ	gpioa_odr_leds,     (0x7 << 5)
.equ	gpiob_odr_leds,     (0x1 << 6)
.equ	gpioa_otyper_leds,  (0x3 << 6) // The first one should not be open-drain.
.equ	gpiob_otyper_leds,  gpiob_odr_leds

.equ	gpioa_leds_count,   3 // How many leds correspond to first gpio - gpioa
.equ	gpioa_leds_lsb_pos, 5 // What is the position of leds lsb in gpioa
.equ	gpiob_leds_lsb_pos, 6 // What is the position of led  lsb in gpiob

.equ	last_led_position,  (0x1 << 3) // Four LEDs total
.equ	init_led_position,  (0x3 << 0) // The inital value for the shift value (1 means off)
.equ	mask_led_position,  0xF // The mask for all leds (4 leds total)

.equ	count_1ms,          7999  // pocet opakovani na 1 ms
.equ	delay,              500   // zpozdeni 500 ms

.section .data
shift_value:
  .word

.section .text

.thumb_func
leds_init:
	push 	{R0-R3};

//  Enable GPIOA in  RCC->AHB1ENR
	ldr		R0, =rcc_reg;
	ldr		R1, =rcc_ahb1enr_offset;
	ldr		R2, [R0, R1];
	ldr		R3, =rcc_gpioa_en;
  orr   R2, R3           ;
	ldr		R3, =rcc_gpiob_en;
	orr		R2, R3;
	str		R2, [R0, R1];

	ldr		R0, =gpio_a;
	ldr		R1, =gpio_moder_offset;
	ldr		R2, =gpioa_moder_leds;
  ldr   R3, [R0, R1]         ;
  orr   R3, R2
	str		R3, [R0, R1];

	ldr		R0, =gpio_b;
	ldr		R1, =gpio_moder_offset;
	ldr		R2, =gpiob_moder_leds;
  ldr   R3, [R0, R1]         ;
  orr   R3, R2
	str		R3, [R0, R1];

//  Output open-drain
	ldr		R0, =gpio_a;
	ldr		R1, =gpio_otyper_offset;
	ldr		R2, =gpioa_otyper_leds;
	str		R2, [R0, R1];

	ldr		R0, =gpio_b;
	ldr		R1, =gpio_otyper_offset;
	ldr		R2, =gpiob_otyper_leds;
	str		R2, [R0, R1];

	pop 	{R0-R3};
	bx		lr;

.thumb_func
shift:
  push {R0-R1}
  ldr R0, =shift_value
  ldr R1, [R0]

  // Shift right by one
  lsrs R1, R1, #1

  // Skip next section if no carry
  bcc shift_next

  // Set last bit, as one has just disappeared
  ldr R0, =last_led_position
  orr R1, R0

shift_next:
  ldr R0, =shift_value
  str R1, [R0]

  bx lr

.thumb_func
write_leds:
  push {R0-R4}
  // GPIOA first
  // Load the shift value, to R1
  ldr R0, =shift_value
  ldr R2, [R0] // R2 contains the shift position
  // Load gpioa leds lsb position
  ldr R0, =gpioa_leds_lsb_pos
  // Shift the shift value to led positions
  lsl R2, R2, R0
  // Mask the value with gpioa_odr_leds
  ldr R0, =gpioa_odr_leds
  and R2, R0

  ldr R0, =gpio_a
  ldr R1, =gpio_odr_offset
  // Store the shift value in R2.
  // NOTE: Ideally, it should only set the few
  // bits, but it does not matter since other
  // bits are always zero in this program.
  str R2, [R0, R1]

  // GPIOB second
  // Load the shift value, to R1
  ldr R0, =shift_value
  ldr R2, [R0] // R2 contains the shift position
  // Shift right by number of leds in gpioa
  ldr R0, =gpioa_leds_count
  lsr R2, R2, R0
  // Shift the shift value to led positions
  ldr R0, =gpiob_leds_lsb_pos
  lsl R2, R2, R0
  // Mask the value with gpiob_odr_leds
  ldr R0, =gpiob_odr_leds
  and R2, R0

  ldr R0, =gpio_b
  ldr R1, =gpio_odr_offset
  // Store the shift value in R2.
  // NOTE: Ideally, it should only set the few
  // bits, but it does not matter since other
  // bits are always zero in this program.
  str R2, [R0, R1]

  pop {R0-R4}
  bx lr

.thumb_func
delay_loop:
	push 	{R0-R1};
	ldr		R1, =delay

//  Load counter for 1ms delay
.thumb_func
load_1ms:
	ldr		R0, =count_1ms

.thumb_func
count_loop:
// Loop through the counter to 1 ms
	subs 	R0, #1
	bne		count_loop

// Loop through the number of ms (delay value)
	subs  R1, #1
	bne		load_1ms

	pop 	{R0-R1};
	bx		lr;

.thumb_func
main:
  bl leds_init;

  ldr R0, =shift_value
  ldr R1, =init_led_position
  str R1, [R0]

main_loop:
  bl shift;
  bl write_leds;
  bl delay_loop;

  b main_loop;