Macro VS Subroutine

[Cisano]

Hi, I have a problem.

I created a macro for update GRP1 X coordinate. Not the positioning, but only the value.

The code is executed during vertical blank time.

Y is used inside the Macro. All works great.

 

ldy #0
UpdateGRP1_Macro

 

I tried to use the same code by a Subroutine.

ldy #0
jsr UpdateGRP1

It doesn't work good. The sprite movement is not smooth.

Is there something I don't know about these 2 way to use code?

Thanks.

 

Edited November 29, 2021 by Cisano

[Thomas Jentzsch]

Can you show some more code? What you describe should not happen.

[Cisano]

        lda FrqFrameGRPX
        and MaskOne,y
        bne .end
        lda XposGRP1,y
        cmp StartOne,y
        bne .other
        lda #1
        sta CursorSpringOne,y
        jmp .opera
.other
        ;lda XposGRP1,y
        cmp ArriveOne,y
        bne .opera
        lda #$ff
        sta CursorSpringOne,y
.opera
        clc
        lda CursorSpringOne,y
        adc XposGRP1,y
        sta XposGRP1,y
.end
        lda StartOne,y
        cmp ArriveOne,y
        bne exit
        sta XposGRP1,y
exit
		rts

The Macro has not the rts.

[Thomas Jentzsch]

Looks fine.

 

Since the code of a macro is included directly inside the code, this changes the following addresses. I suppose a page penalty is happening now.

[Cisano]

Penalty? You mean extra cycles?

[stepho]

Just some generic things to look for...

 

1.

Jumping to the subroutine takes a few cycles.
Returning also takes a few cycles.

You said it is in the vertical blank code but is it time sensitive?

 

2.

Storing the return address takes a few bytes of stack.

Is the stack big enough?

 

3.

Some bank switching schemes trigger off subroutine calls (ie when you write/read the return address in the first few elements of the stack).

[Cisano]

1. I set a timer and the vertical blank time is not all used.

2. yes it is

3. no bank switching for now

 

I just moved the subroutine in another part of the cose, upper, and it works good.

It is the normality or there are solutions?

Thanks.

Edited November 30, 2021 by Cisano

[Thomas Jentzsch]

15 hours ago, Cisano said:

Penalty? You mean extra cycles?

Yup.

 

Always when a branch or an indexed table access crosses a 256 bytes page the instruction takes one extra cycle.

To check this during assembly I have written a simple macro:

  MAC CHECKPAGE
    IF >. != >{1}
      ECHO ""
      ECHO "ERROR: different pages! (", {1}, ",", ., ")"
      ECHO ""
      ERR
    ENDIF
  ENDM

If your branch moves backward, put this behind the branch and provide the branch address. If it is a table, put it at the end of the table and provide the table address.

Edited November 30, 2021 by Thomas Jentzsch