73 lines
1.5 KiB
NASM
73 lines
1.5 KiB
NASM
;; sketching around subroutines with return stack
|
||
|
||
; I think the sketching below contains some useful ideas,
|
||
; but the new `*ADDR` assembly op should make this a whole lot simpler
|
||
|
||
; sketches follow ..........
|
||
|
||
; jump table
|
||
; ----------
|
||
; each function has a label
|
||
; and you can get to it with `JMP @label`
|
||
; ...but you can't push a label onto a stack
|
||
; so instead we give each function a number
|
||
; and stash that number in $24 when jumping to a subroutine
|
||
; then when the subroutine ends, it jumps here
|
||
; where that number gets mapped back onto the function's label
|
||
|
||
=*reentryPoint
|
||
=routine1 $1
|
||
|
||
lda =routine1
|
||
sto $19 ; contains ID # for the next fn to jump to
|
||
|
||
@jump_table
|
||
hop $1
|
||
jmp @jt2
|
||
jmp @example_computation
|
||
|
||
@jt2
|
||
hop $2
|
||
; jmp @jt3
|
||
nop
|
||
; etc …
|
||
|
||
jmp @end
|
||
|
||
@example_computation
|
||
lda 5
|
||
sto $20
|
||
lda 3
|
||
sto $21
|
||
; $19 still has the # for this routine
|
||
; but let’s pretend it doesn’t and demonstrate updating it
|
||
lda $1
|
||
sto $19
|
||
jmp @greater?
|
||
|
||
; call with numbers to test in $20 and $21
|
||
; result is stored in acc
|
||
@greater?
|
||
; lda ($20)
|
||
; sub ($21)
|
||
; todo…
|
||
; wouldn’t it be great to have a “hop if neg” op…
|
||
; do we have to just subtract numbers until we get 0?
|
||
|
||
; no!
|
||
; here’s an approach that’s at least better than that
|
||
lda ($21)
|
||
sto $22 ; stash
|
||
@loop
|
||
lda ($21)
|
||
sub $1
|
||
sto $22 ; stash
|
||
sub ($20)
|
||
hop $0
|
||
jmp @loop
|
||
sto $1
|
||
jmp $jmp_table
|
||
; ok this isn’t quite it… we also need to chexk if we hit 0 by just deceementinf and if so retuen 0
|
||
|
||
@end
|
||
END |