SETQ - Quick question for Chip

Cluso99

Chip,
Just wanted to confirm this is OK to do...

  if_e    setq    #len1-1
  if_ne   setq    #len2-1
          rdlong  cog0,hubptr

ie the setq may not be immediately before the rdlong

evanh

I think the first SETQ will never reach the RDLONG. From memory, even a NOP clears it. Same for ALTx instructions.

Cluso99

May as well also ask, can setq be set more than once and will the new setting overwrite the previous?

evanh

Cluso99 wrote: »

... will the new setting overwrite the previous?

More accurately, it will be rearmed because the first SETQ value is already consumed.

cgracey

If the 2nd SETQ gets cancelled, it looks like a NOP, making the RDLONG just read a single long, as if no SETQ was used.

Cluso99

So the SETQ must immediately precede the RD/WR/LONG.
So no non-executed instructions can intervene.

ozpropdev

Instead of

  if_e    setq    #len1-1
  if_ne   setq    #len2-1
          rdlong  cog0,hubptr

you could use

  if_e    altd    zero,#len1-1
          setq    #len2-1
          rdlong  cog0,hubptr

where "zero" is a cog register containg #0

Cluso99

I had to settle for

        if_ne   mov     reply,            #cog_len-1    ' set to copy whole cog (496 longs)
        if_e    mov     reply,            #cog_len80-1  ' set to copy 512 bytes (128 longs)       
                setq    reply                           ' length -1
                rdlong  cog_start0,       _hubdata      ' copy loaded code into cog
        if_ne   jmp     #$0                             ' NZ: execute loaded cog code
                jmp     #$020                           ' Z:  execute loaded cog code from $080+