#ifndef __DELAY_INC_ASM_INCLUDED__
#define __DELAY_INC_ASM_INCLUDED__

.equ TAKTE_PRO_MS = (F_CPU / 1000)
.equ TAKTE_PRO_NS = (F_CPU / 1000000)
.equ TAKTE_PRO_SCHLEIFE = 6
.equ SCHLEIFEN_PRO_MS = (TAKTE_PRO_MS / TAKTE_PRO_SCHLEIFE)
.equ SCHLEIFEN_PRO_NS = (TAKTE_PRO_NS / TAKTE_PRO_SCHLEIFE)
.if (SCHLEIFEN_PRO_MS>0xffff) 
  .error "Aktuelle Taktfrequenz ergibt mehr als 0xffff Schleifendurchläufe/ms!"
.endif

.macro M_DELAY_MICROSECOND_IN_LOOP
; ein Schleifendurchlauf braucht 3 Takte.
; Bei 3 MHz erzeugt dieses Makro nichts.
.if (F_CPU > 3000000)
	nop ; Bei 4 MHz einen NOP 
.endif
.if (F_CPU > 4000000)
	nop ; bei 5 MHz zwei NOPs 
.endif
.if (F_CPU > 5000000)
	nop ; usw..
.endif
.if (F_CPU > 6000000)
	nop
.endif
.if (F_CPU > 7000000)
	nop
.endif
.if (F_CPU > 8000000)
	nop
.endif
.if (F_CPU > 9000000)
	nop
.endif
.if (F_CPU > 10000000)
	nop
.endif
.if (F_CPU > 11000000)
	nop
.endif
.if (F_CPU > 12000000)
	nop
.endif
.if (F_CPU > 13000000)
	nop
.endif
.if (F_CPU > 14000000)
	nop
.endif
.if (F_CPU > 15000000)
	nop
.endif
.if (F_CPU > 16000000)
	nop
.endif
.if (F_CPU > 17000000)
	nop
.endif
.if (F_CPU > 18000000)
	nop
.endif
.if (F_CPU > 19000000)
	nop
.endif
.endm

.macro M_DELAY_MICROSECOND
	nop
.if (F_CPU > 1000000)
	nop
.endif
.if (F_CPU > 2000000)
	nop
.endif
.if (F_CPU > 3000000)
	nop
.endif
.if (F_CPU > 4000000)
	nop
.endif
.if (F_CPU > 5000000)
	nop
.endif
.if (F_CPU > 6000000)
	nop
.endif
.if (F_CPU > 7000000)
	nop
.endif
.if (F_CPU > 8000000)
	nop
.endif
.if (F_CPU > 9000000)
	nop
.endif
.if (F_CPU > 10000000)
	nop
.endif
.if (F_CPU > 11000000)
	nop
.endif
.if (F_CPU > 12000000)
	nop
.endif
.if (F_CPU > 13000000)
	nop
.endif
.if (F_CPU > 14000000)
	nop
.endif
.if (F_CPU > 15000000)
	nop
.endif
.if (F_CPU > 16000000)
	nop
.endif
.if (F_CPU > 17000000)
	nop
.endif
.if (F_CPU > 18000000)
	nop
.endif
.if (F_CPU > 19000000)
	nop
.endif
.endm

#ifdef USE_DELAY_US
Delay_us:
; wartet r16 µs lang
; Zusätliche Takte, die abgezogen werden müssen:
; 3 (rcall) + 4 (ret) + 8 (push und pop) 2 (in und out) + 1 (subi) + 1 (brne ohne branch)
; => 17 Takte
.equ Takte_pro_us = F_CPU/1000000
.equ DelayMicrosecondsAbzug = (17 / Takte_pro_us)
	push r16 ; 2 Takte
	push r17 ; 2 Takte
	in r17,SREG
	subi r16, DelayMicrosecondsAbzug ; 1 Takt
DelayMicrosecondsLoop:
	M_DELAY_MICROSECOND_IN_LOOP
	dec r16
	brne DelayMicrosecondsLoop ; 1 Takt
	out SREG, r17
	pop r17 ; 2 Takte
	pop r16 ; 2 Takte
	ret ; 4 Takte
#endif

#if (defined(USE_DELAY_MS) || defined(USE_DELAY_10MS) || defined(USE_DELAY_100MS) || defined(USE_DELAY_1S))
Delay_ms:
	push r16
	push r17
	push r18
	push r19
	push r20
	ldi r19,0
	ldi r20,1
Delay_ms_S_ms:
	ldi r17,LOW(SCHLEIFEN_PRO_MS)
	ldi r18,HIGH(SCHLEIFEN_PRO_MS)
Delay_ms_S:
	sub r17,r20 ; 1 Takt
	sbc r18,r19 ; 1 Takt
	cp r17,r19 ; 1 Takt
	cpc r18,r19 ; 1 Takt
	brne Delay_ms_S ; 2 Takte
	dec r16
	brne Delay_ms_S_ms
	pop r20
	pop r19
	pop r18
	pop r17
	pop r16
	ret
#endif

#ifdef USE_DELAY_10MS
Delay_10ms:
	push r16
	ldi r16,10
	rcall Delay_ms
	pop r16
#endif

#ifdef USE_DELAY_100MS
Delay_100ms:
	push r16
	ldi r16,100
	rcall Delay_ms
	pop r16
#endif
 
#ifdef USE_DELAY_1S
Delay_1s:
	push r16
	ldi r16,200
	rcall Delay_ms
	rcall Delay_ms
	rcall Delay_ms
	rcall Delay_ms
	rcall Delay_ms
	pop r16
#endif

/* Alte Version:
Delay_us:
	push r16
	push r17
	push r18
	push r19
	push r20
	ldi r19,0
	ldi r20,1
Delay_us_S_us:
	ldi r17,LOW(SCHLEIFEN_PRO_NS)
	ldi r18,HIGH(SCHLEIFEN_PRO_NS)
Delay_us_S:
	sub r17,r20 ; 1 Takt
	sbc r18,r19 ; 1 Takt
	cp r17,r19 ; 1 Takt
	cpc r18,r19 ; 1 Takt
	brne Delay_us_S ; 2 Takte
	dec r16
	brne Delay_us_S_us
	pop r20
	pop r19
	pop r18
	pop r17
	pop r16
	ret
*/

#endif ; #ifndef __DELAY_INC_ASM_INCLUDED__