Cameron Kaiser wrote: I see, the 6502 draws a LOT of attention, mine too ... may be too late? Anyhow, having just revived my memories with a C=64 from ebay and looking into LUnix for it :-) I've been thinking about how to make stack frames on the 6502. Clearly with a stack between $0100 and $01ff it is a bit tight. But, if the 65CE02 should have a 16 bit SP, what should be the problem with putting arguments on the stack? Why do you need to move the return address? You don't pull the arguments, you LDA them! That's what the TSX opcode must be for after all, the X register acting as some sort of base pointer! caller: LDA argN PHA ... LDA arg1 PHA JSR callee TSX INX N TXS callee: TSX LDA $0101+1,X ; for arg1 ... LDA $0101+N,X ; for argN ... RTS See, no PLA involved at all. In fact as I remember the PHA and PLA opcodes where among the most cycle expensive you could have. So, one could even speed up the argument placement: caller: TSX DEX N TXS LDA arg1 STA $00FF+1,X ... LDA argN STA $00FF+N,X JSR callee TSX INX N TXS Just to challenge the 8-bit SP a little, let's put local variables onto the stack :-) callee: TSX DEX M ; reserve space for M local variables TXS LDA $00FF+1,X ; for loc1 ... LDA $00FF+M,X ; for locM ... LDA $0101+M+1,X; for arg1 ... LDA $0101+M+N,X; for argN ... TSX INX M TXS RTS Now we just need some frame-pointer mechainsm on the zero page and we can do everything the ix86 can do :-) It took me so many years to find out why the TSX makes sense :-) regards -Gunther -- Gunther Schadow, M.D., Ph.D. gschadow(a)regenstrief.org Medical Information Scientist Regenstrief Institute for Health Care Adjunct Assistent Professor Indiana University School of Medicine tel:1(317)630-7960 http://aurora.regenstrief.org