something strange is happening

const.h

@@ -0,0 +1,4 @@
+#pragma once
+
+#define VERTICAL_OFFSET 30
+

main.S

@@ -1,25 +1,27 @@
 #include <avr/io.h>
 #include "macro.h"
+#include "const.h"
 
 .data
+vertical_offset:
+.byte 0
 
-status:
+frame:
-.byte 0x0
+.byte 0
-
-offset:
-.byte 78
 
 line:
-.word 0
+.byte 0
 
 .text
 
-.global main_s
+.global main
-main_s:
+main:
-    ldi r16, 0x30
+    ldi r16, 0x30       ; port B, pin 4 and 5 as output
     sts DDRB, r16
 
-    ldi r16, (1 << IVCE)    ; set vector at address 0x0
+    ; set interrupt vectors at address 0x0, not bootloader
+    ; timing is important, see atmel datasheet
+    ldi r16, (1 << IVCE)
     ldi r17, 0
     out IO(MCUCR), r16
     out IO(MCUCR), r17
@@ -30,44 +32,25 @@ main_s:
     ldi r16, 0x3        ; external interrupt 0 and 1, mask enable
     sts EIMSK, r16
 
-    sei
+    sei                 ; global interrupt enable
-
-again:
-    sbi IO(PORTB), 5
-    ldi r16, 0x1
-    sts status, r16
-
-wait1:
-    ldi r16, 0xff
-    ldi r17, 0xff
-    ldi r18, 0xff
 
 1:
-    dec r16
+    sbi IO(PORTB), 5
-    brne 1b
+    ldi r18, 0
-    dec r17
+2:
-    brne 1b
     dec r18
-    brne 1b
+    brne 2b
 
     cbi IO(PORTB), 5
-    ldi r16, 0x0
-    sts status, r16
 
-wait2:
+    ldi r18, 0
-    ldi r16, 0xff
+3:
-    ldi r17, 0xff
-    ldi r18, 0xff
-
-1:
-    dec r16
-    brne 1b
-    dec r17
-    brne 1b
     dec r18
-    brne 1b
+    brne 3b
 
-    jmp again
+    rjmp 1b
+
+    jmp .
 
 .global int_horizontal_sync
 int_horizontal_sync:
@@ -75,26 +58,34 @@ int_horizontal_sync:
     in r31, IO(SREG)    ; status register
     push r31
 
+#if 0
+    lds r31, vertical_offset
+    cpi r31, 0
+    breq enter1
+    dec r31
+    sts vertical_offset, r31
+    jmp int_horizontal_sync_end
+
+enter1:
     lds r31, line
-    cpi r31, 255
+    cpi r31, 0
-    brsh int_horizontal_sync_end
+    breq int_horizontal_sync_end
     inc r31
     sts line, r31
 
-    lds r31, status
+    lds r31, frame
-    cpi r31, 0
+    lsr r31
-    brne int_horizontal_sync_end
+    lsr r31
-
+    breq 2f
-    lds r31, offset
+1:
     dec r31
-    cpi r31, 1
+    brne 1b
-    brne 1f
+#endif
+
     ldi r31, 140
 1:
-    sts offset, r31
-loop:
     dec r31
-    brne loop
+    brne 1b
 
     sbi IO(PORTB), 4
     nop
@@ -118,10 +109,22 @@ int_horizontal_sync_end:
 .global int_vertical_sync
 int_vertical_sync:
     push r31
-    ldi r31, 0
+    in r31, IO(SREG)
+    push r31
+
+    lds r31, frame
+    inc r31
+    sts frame, r31
+
+    ldi r31, 1
     sts line, r31
-    ldi r31, 140
+
-    sts offset, r31
+    ldi r31, VERTICAL_OFFSET
+    sts vertical_offset, r31
+
+int_vertical_sync_end:
+    pop r31
+    out IO(SREG), r31
     pop r31
     reti
 

main.c

@@ -1,9 +1,6 @@
 #include <avr/io.h>
 #include <avr/interrupt.h>
 
-void main_s(void);
-
-
 ISR(INT0_vect, ISR_NAKED) {
     asm("jmp int_vertical_sync");
 }
@@ -11,6 +8,3 @@ ISR(INT1_vect, ISR_NAKED) {
     asm("jmp int_horizontal_sync");
 }
 
-int main() {
-    main_s();
-}