Question about Kernal function GETIN $FFE4

KingTrode · Post by **KingTrode** » Fri May 14, 2010 3:20 am

According to the PRG registers affected by this are A and X.

I seem to be getting corruption of Y register after calling this.

Is the PRG incorrect?

Kananga · Post by **Kananga** » Fri May 14, 2010 4:28 am

I'd always assume that kernel functions use all registers. There are not many registers in a 6502 anyway.
You could have a look into the ROM listing or step into the routine using a debugger to see what's happening.

Diddl · Post by **Diddl** » Fri May 14, 2010 4:49 am

Kananga wrote:I'd always assume that kernel functions use all registers. There are not many registers in a 6502 anyway.
You could have a look into the ROM listing or step into the routine using a debugger to see what's happening.

yes!

In case of GETIN function:

If GETIN uses keyboard:

- if no key pressed (empty key buffer) only AC ist used
- if key in buffer, all registers are lost

If other sources (IEC) register are lost.

KingTrode · Post by **KingTrode** » Fri May 14, 2010 4:54 am

Kananga wrote:I'd always assume that kernel functions use all registers. There are not many registers in a 6502 anyway.
You could have a look into the ROM listing or step into the routine using a debugger to see what's happening.

Most kernel functions will most likely use all registers, it's which ones are preserved by the function and restored on return (other than those which are used to return values)

CHROUT $FFD2 affects none for example, otherwise a print loop would involve some form of --> TXA/TYA PHA before and PLA TAY/TAX afterwards.

It's not a big problem as I'll just push it on the stack, was just wondering if this is incorrect - and how many other errors might be in the book.

KingTrode · Post by **KingTrode** » Fri May 14, 2010 5:10 am

Diddl wrote:
Kananga wrote:I'd always assume that kernel functions use all registers. There are not many registers in a 6502 anyway.
You could have a look into the ROM listing or step into the routine using a debugger to see what's happening.
yes!

In case of GETIN function:

If GETIN uses keyboard:

- if no key pressed (empty key buffer) only AC ist used
- if key in buffer, all registers are lost

If other sources (IEC) register are lost.

So the PRG is incorrect then.

Thanks for all replies

Kananga · Post by **Kananga** » Fri May 14, 2010 11:57 am

KingTrode wrote: So the PRG is incorrect then.

That's why I always have a commented ROM listing on my desk when programming for the VIC.

FD22 · Post by **FD22** » Sun May 16, 2010 4:57 am

Kananga wrote:
KingTrode wrote: So the PRG is incorrect then.
That's why I always have a commented ROM listing on my desk when programming for the VIC.

I have an annotated memory map layout which includes the names of the ROM routine entry points, but doesn't include comments about their register dependencies and usage - if you have that, do you feel like sharing? I'll post my VICMap file in exchange...

Kananga · Post by **Kananga** » Sun May 16, 2010 11:15 am

FD22 wrote: I have an annotated memory map layout which includes the names of the ROM routine entry points, but doesn't include comments about their register dependencies and usage - if you have that, do you feel like sharing? I'll post my VICMap file in exchange...

Is your map the one found on funet/zimmers?
The commented ROM listing I have is in the German book "VC-20 Intern" (published by Data Becker, 1983). I don't know if anybody scanned it already and it is in German, of course.

FD22 · Post by **FD22** » Sun May 16, 2010 2:41 pm

Nope - well, not exactly, it's obviously the VIC memory map, but I've annotated every writeable memory location and named every ROM routine (using either the commonly-accepted name, or a representative one of my own making if there didn't seem to be one).

Does anyone have a link to an annotated / commented ROM disassembly?

Kananga · Post by **Kananga** » Mon May 17, 2010 1:05 pm

I just found this in The Complete Commodore Inner Space Anthology (ftp.zimmers.net/pub/cbm/manuals/anthology/index.html):

GETIN FFE4 "Get character from current input device"
output: .A .. data .X .. alt. .Y alt.

(Machine Language Section: BASIC 2.0/4.0 Kernal Routines, VIC 20/Commodore 64 Kernal Routines)

Not a ROM listing, but it looks like a useful reference.

Kananga · Post by **Kananga** » Mon May 17, 2010 1:18 pm

FD22 wrote:Nope - well, not exactly, it's obviously the VIC memory map, but I've annotated every writeable memory location and named every ROM routine (using either the commonly-accepted name, or a representative one of my own making if there didn't seem to be one).

Sounds like your map is more comprehensive than the one I use for my debugger (http://vic20emu.googlecode.com/svn/trun ... vicrom.sym) (It's the one from zimmers.net)
Would you like to share your map?

FD22 · Post by **FD22** » Mon May 17, 2010 3:16 pm

Here is a link to my VICMap file, which is designed to be used as an 'include' for DASM-based VIC Assembler projects (which is what I created it for). You can rename it as '.asm' if you're using that as your Assembler extension - I already have that associated with an x86 dev-tool, so I use .6502 to map DASM source files to a different toolchain.

Here it is: http://vicdev.googlecode.com/files/VICMap.6502

Kananga · Post by **Kananga** » Tue May 18, 2010 12:05 am

Thanks for sharing!
I'd like to borrow from your map what's missing from my symbol file for the Debugger, if you don't mind.

Diddl · Post by **Diddl** » Tue May 18, 2010 2:17 am

Thanks for that definition file.

FD22 · Post by **FD22** » Tue May 18, 2010 11:05 am

If I minded anyone using it, or a derivative of it, I wouldn't have posted it.

I've gleaned an immense amount of VIC-related information from this forum and other internet resources, so I consider posting stuff like VICMap a fair return for all the knowledge now in my head that other peoples' work has contributed to.