Since I've been running into a bunch of nitpicking questions about the 65816 during my attempt to emulate the chip, I thought I might keep them in one thread for future intrepid coders instead of spreading them out all over the forum like I did with the TXS/SP question (viewtopic.php?f=4&t=3446). I'm assuming that the same problems will come up again.

Here's my latest one: Is there anything to prevent me from manipulating the Direct Register (D) in emulation mode, and what effect would that have? The Programming Manual just says that it gets set to 0 when we switch, but there doesn't seem to be anything that keeps me from doing a TCD in emulation mode, and the entry for Direct Page Addressing (p. 298) doesn't say anything about "65816/65802 only". I've checked BDD's 65C816 Programming Tips entry for DP (viewtopic.php?f=2&t=3198&p=36677e#p36677), but nothing there either, it all seems to be about native mode.

So if I switch to emulated mode, change D to, say, $1000, and then do some ZP access at $10, will the processor just ignore D or actually give me $1010?

Obviously, this would be a weird thing to do, but you never know.

The general rule is that all instructions work in both modes, although the results may differ between the two modes. In emulation mode, a sequence such as:

Code:

         lda #$c8
         xba
         lda #$40
         tcd

will set DP to $C840. On the other hand:

Code:

         lda #$c8
         xba
         lda #$40
         tcs

will not set SP to $C840, as the MSB of the stack pointer is hard-wired to $01 in emulation mode. Instead, SP will be $0140.

As a reminder, direct page and stack accesses always occur in bank $00, and the JMP (<addr>) instruction is effectively JMP ($00<addr>).

_________________
x86?  We ain't got no x86.  We don't NEED no stinking x86!

So with XBA, TCD and TDC, emulation mode has access to a further two bytes of register?

Correct.

XBA is particularly useful in emulation mode, as you effectively have two accumulators at your disposal. It's just a matter of keeping track of which accumulator is holding what, and remembering that a PHA only pushes the A-accumulator.

Even in native mode, XBA is very useful. Consider, for example, a SCSI driver that is setting up a command descriptor block (CDB). Data transfer sizes in CDBs are written in big-endian format, but the 65C816 processes in little-endian. So if a data transfer size has been computed and is presently loaded in the 16-bit accumulator, it will be in little-endian format. Prior to writing the size value into the CDB, an XBA would reverse endianess.

Interesting aside: XBA counter-intuitively affects the N and Z flags in SR according to the value transferred into the A-accumulator, regardless of the current m bit setting in SR. Hence the following code would set N and clear Z:

Code:

         rep #%00100000        ;16 bit accumulator
         lda #$8000            ;.C = $8000
         xba                   ;.C = $0080
         brk

On the other hand:

Code:

         rep #%00100000        ;16 bit accumulator
         lda #$8000            ;.C = $8000
         and #$ff00            ;.C = $8000
         brk

would also set N and clear Z, as all arithmetic and Boolean operations on the 16 bit accumulator condition N according to the value of bit 15, not bit 7.

_________________
x86?  We ain't got no x86.  We don't NEED no stinking x86!

So if I'm willing to force the issue by manipulating D in emulation mode, I can put the Direct Page anywhere in the first 64k in that mode as well? That of course simplifies the code for the direct page addressing mode, but sounds like an accident waiting to happen in emulation mode ... thanks!

The general rule is just that, so I'd be cautious about reading too much detail beween the lines. This is an esoteric question, and the answer is probably worth confirming.

Sometimes an actual experiment is required. (BDD? Garth? Anyone?) Thanks. I think Scot would prefer a definite answer if one is available. The question is,

_________________
In 1988 my 65C02 got six new registers and 44 new full-speed instructions!
https://laughtonelectronics.com/Arcana/ ... mmary.html

That's what both the data sheet and the Lichty and Eyes programming manual say. I've not tried it.

_________________
x86?  We ain't got no x86.  We don't NEED no stinking x86!

It's when you guys use words like "esoteric" and "experimental" that I start to wonder if I'm in over my head . Thanks - for the moment, I'll go with what we think happens and put a note in the TODO and MANUALs with the background until somebody can test this.

It would be interesting to have a list of forum members who have a working '816 system to experiment with, and which of them is willing to help out.

I believe Bruce has a IIGS. I've had the beeb816 but nothing I can run with at present. BDD has the POC series.

I could try some tests on my SXB - Its one of the reasons I bought it as some of the instruction descriptions weren't clear enough for me to write the PIC24 emulation code for them.

_________________
Andrew Jacobs
6502 & PIC Stuff - http://www.obelisk.me.uk/
Cross-Platform 6502/65C02/65816 Macro Assembler - http://www.obelisk.me.uk/dev65/
Open Source Projects - https://github.com/andrew-jacobs

That'd be great, Andrew!

To be clear, neither outcome will surprise me. I posted simply to point out that doc has limitations. As with test equipment, it's constantly necessary to question the reliability of the info you're collecting because the info gets more and more questionable as the limitations are approached.

Re: the answer to Scot's question, I haven't yet seen a reference that's specific rather than merely suggestive. To me that says we've reached the limit -- the doc simply doesn't answer the question satisfactorily.

_________________
In 1988 my 65C02 got six new registers and 44 new full-speed instructions!
https://laughtonelectronics.com/Arcana/ ... mmary.html

My guess is that you can write to D in emulation mode, but the value won't affect any zero page addressing. Edit: did a little reading, and it seems my guess is wrong!

I'd have to burn a special ROM to make the POC unit boot into emulation mode or write some code to temporarily run it in emulation mode. One more thing on the bucket list.

_________________
x86?  We ain't got no x86.  We don't NEED no stinking x86!

The designers of the '816 thoughtfully provided an XCE instruction to help you.

If I didn't make any coding errors, executing this small test will cause location $1000 to contain $cd if DPR is ignored in emulation mode, and $ab if it is not.

Mike B.

[Edit: Fixed obvious and heinous coding error.]

Sorry I missed that, Mike! But I won't tease you for being too quick to hit the Submit button -- I'm often guilty of that myself. Thanks for posting.

_________________
In 1988 my 65C02 got six new registers and 44 new full-speed instructions!
https://laughtonelectronics.com/Arcana/ ... mmary.html