Copper SKIP after SKIP

[neoman]

Hello,

I'm playing around with Copper SKIPs and JUMPs and I've noticed that a SKIP directly after a successful SKIP is executed though. Copper code is something like:

MOVE
SKIP - this one is successful, so it should skip over the next SKIP to JUMP, but he continue just right ....
SKIP - here... :-P
JUMP

copper.cpp, line 6318:

if (cop_state.i1 & 1) { // WAIT or SKIP
	cop_state.ignore_next = 0;
	if (cop_state.i2 & 1)
		cop_state.state = COP_skip_in2;
	else
		cop_state.state = COP_wait_in2;
#ifdef DEBUGGER
	if (debug_dma)
		record_dma (0x8c, cop_state.i2, cop_state.ip - 2, old_hpos, vpos, DMARECORD_COPPER);
	if (memwatch_enabled)
		debug_wgetpeekdma_chipram(cop_state.ip - 2, cop_state.i2, MW_MASK_COPPER, 0x8c);
#endif
}

Here ignore_next is reset to 0 instead of emitting a 0x1FE/NO-OP like in the MOVE command.

Is that intended and can I find it in the hardware docs?

Thanks,
neoman

[Toni Wilen]

It is correct and matches real hardware.

SKIP does not actually skip anything.

It only sets internal variable that causes following MOVE (only if following instruction is MOVE) to do the move to dummy address. MOVE cycle sequence is still the same, "skipped" or not.

Nothing happens if following instruction is another SKIP or WAIT.

[neoman]

Okay, I see. So no way to do negative SKIPs =/

Thanks!

[neoman]

Okay, I solved my problems with a sequence like

SKIP
JUMP1
JUMP2

That's what I meant by "negative" SKIP. Basically I needed if then else and this is working fine now =) Thanks again.