2000-05-28 13:40:48 +00:00
|
|
|
;
|
2009-08-17 20:48:28 +00:00
|
|
|
; Ullrich von Bassewitz, 2009-08-17
|
2000-05-28 13:40:48 +00:00
|
|
|
;
|
|
|
|
; CC65 runtime: multiplication for ints
|
|
|
|
;
|
|
|
|
|
2013-05-09 11:56:54 +00:00
|
|
|
.export tosumulax, tosmulax
|
2009-08-17 20:48:28 +00:00
|
|
|
.import mul8x16, mul8x16a ; in mul8.s
|
2013-05-09 11:56:54 +00:00
|
|
|
.import popsreg
|
|
|
|
.importzp sreg, tmp1, ptr4
|
2009-08-17 20:48:28 +00:00
|
|
|
|
|
|
|
|
|
|
|
;---------------------------------------------------------------------------
|
|
|
|
; 16x16 multiplication routine
|
2000-05-28 13:40:48 +00:00
|
|
|
|
|
|
|
tosmulax:
|
|
|
|
tosumulax:
|
2013-05-09 11:56:54 +00:00
|
|
|
sta ptr4
|
2009-08-17 20:48:28 +00:00
|
|
|
txa ; High byte zero
|
|
|
|
beq @L3 ; Do 8x16 multiplication if high byte zero
|
2013-05-09 11:56:54 +00:00
|
|
|
stx ptr4+1 ; Save right operand
|
|
|
|
jsr popsreg ; Get left operand
|
2009-08-17 20:48:28 +00:00
|
|
|
|
|
|
|
; Do ptr4:ptr4+1 * sreg:sreg+1 --> AX
|
|
|
|
|
2013-05-09 11:56:54 +00:00
|
|
|
lda #0
|
|
|
|
ldx sreg+1 ; Get high byte into register for speed
|
2009-08-17 20:48:28 +00:00
|
|
|
beq @L4 ; -> we can do 8x16 after swap
|
2013-05-09 11:56:54 +00:00
|
|
|
sta tmp1
|
|
|
|
ldy #16 ; Number of bits
|
2009-08-17 20:48:28 +00:00
|
|
|
|
|
|
|
lsr ptr4+1
|
|
|
|
ror ptr4 ; Get first bit into carry
|
|
|
|
@L0: bcc @L1
|
|
|
|
|
2013-05-09 11:56:54 +00:00
|
|
|
clc
|
|
|
|
adc sreg
|
|
|
|
pha
|
|
|
|
txa ; hi byte of left op
|
|
|
|
adc tmp1
|
|
|
|
sta tmp1
|
|
|
|
pla
|
2009-08-17 20:48:28 +00:00
|
|
|
|
|
|
|
@L1: ror tmp1
|
2013-05-09 11:56:54 +00:00
|
|
|
ror a
|
|
|
|
ror ptr4+1
|
|
|
|
ror ptr4
|
2009-08-17 20:48:28 +00:00
|
|
|
dey
|
|
|
|
bne @L0
|
|
|
|
|
2013-05-09 11:56:54 +00:00
|
|
|
lda ptr4 ; Load the result
|
|
|
|
ldx ptr4+1
|
|
|
|
rts ; Done
|
2009-08-17 20:48:28 +00:00
|
|
|
|
|
|
|
; High byte of rhs is zero, jump to the 8x16 routine instead
|
|
|
|
|
|
|
|
@L3: jmp mul8x16
|
|
|
|
|
|
|
|
; If the high byte of rhs is zero, swap the operands and use the 8x16
|
|
|
|
; routine. On entry, A and X are zero
|
|
|
|
|
|
|
|
@L4: ldy sreg ; Save right operand (8 bit)
|
|
|
|
ldx ptr4 ; Copy left 16 bit operand to right
|
|
|
|
stx sreg
|
|
|
|
ldx ptr4+1 ; Don't store, this is done later
|
|
|
|
sty ptr4 ; Copy low 8 bit of right op to left
|
|
|
|
ldy #8
|
|
|
|
jmp mul8x16a
|
2000-05-28 13:40:48 +00:00
|
|
|
|