mist64 / c64ref

Ultimate Commodore 64 Reference: ROM Disassembly, Memory Map, ...

Ultimate Commodore
64 Reference Guide

An effort to collect C64 reference material
in machine readable form and maintain
scripts to present this material to the web.

   

Material

Reference material collected in this repository.

ROM Disassembly

KERNAL	BASIC	Lang	By / From
◯	⬤	🇺🇸	Microsoft
⬤	◯	🇺🇸	Commodore
⬤	⬤	🇺🇸	Lee Davison
◯	⬤	🇺🇸	Bob Sander-Cederlof
⬤	⬤	🇩🇪	Commodore-64-intern-Buch
⬤	◯	🇺🇸	Magnus Nyman
⬤	⬤	🇺🇸	Marko Mäkelä

Memory Map

Title	Lang	By / From
Mapping the Commodore 64	🇺🇸	Sheldon Leemon
Memory Map mit Wandervorschlägen	🇩🇪	Dr. H. Hauck
Commodore-64-intern-Buch	🇩🇪	Commodore
Reference	🇺🇸	Joe Forster / STA
Comments in the original source	🇺🇸	Microsoft / Commodore
C64 Programmer's Reference Guide	🇺🇸
64map	🇺🇸
Reference	🇺🇸	Jim Butterfield

KERNAL API

Title	Lang	By / From
Commodore 64 Programmer's Reference Guide	🇺🇸	Commodore
COMPUTE!'s VIC-20 and Commodore 64 Tool Kit: Kernal	🇺🇸	Dan Heeb
Machine Language Routines for the Commodore 64 and 128	🇺🇸	Todd D Heimarck Patrick Parrish
Mapping the Commodore 64	🇺🇸	Sheldon Leemon
Commodore 128 intern	🇺🇸	Jörg Schieb Frank Thrun Heinz Wrobel
The almost completely commented C64 ROM disassembly	🇺🇸	Lee Davison
Cracking The Kernal	🇺🇸	Peter Marcotty
Kernal 64 / 128	🇺🇸	Craig Taylor
Commodore 64 standard KERNAL functions	🇺🇸	Joe Forster / STA
C64 KERNAL jump table	🇺🇸	Frank Kontros
Das neue Commodore-64-intern-Buch	🇩🇪	Baloui Brückmann Englisch Felt Gelfand Gerits Krsnik

Contributions

Extensions , Translations , Corrections
of typos and content, .. are welcome!

Credits

The original effort of collecting, converting, formatting
and editing the collected files was done by Michael Steil.

c64ref's Issues

char set render issue in Chrome on Windows

On Windows (but not android), in Chrome (but not in other browsers, such as firefox or duckduckgo), some of the characters are shifted a bit to the right. See image below:

SQRHLF

The fully commented ROM for MS has the SQRHLF bytes starting at B9D5 which is $20.
The $20 at B9D5 is actually the 5th byte of the previous float 2.8853900728 in LOGCN2.
SQRHLF's 5 float bytes start at B9D6.

The incorrect listing is below for context:
.:B9D1 82 38 AA 3B 20 202 ; 2.8853900728 <=== B9D1
070 <=== B9D2
252 <=== B9D3
073 <=== B9D4
040> <=== is actually B9D5 20 (hex)
.:B9D5 80 35 04 F3 34 SQRHLF: 200 ; SQR(0.5) <=== SQRHLF starts at B9D6 80 (200octal)
065
004
363
IFN ADDPRC,<064>
.:B9DB 81 35 04 F3 34 SQRTWO: 201 ; SQR(2.0) <=== is crrect address and label

PETSCII code 255 maps to wrong character

The PETSCII code 255 should be the same character as code 126, but it erroneously maps to code 191 in your PETSCII table. (Perhaps the problem is around here in generate.py.)

For reference:

Link rot pagetable.com

All links to pagetable.com subpages, even the same-domain links found in blog posts there, are dead and just show the latest blog posts.

Also show "real" address (i.e. vector destination)

In addition to the API address (the jump-table), could you also list the real address too, i.e. where the jump-table jumps to. For example, see: https://sta.c64.org/cbm64krnfunc.html This is useful for skipping a layer of indirection and improving speed, where desirable.

PETSCII conversion to Screen Code.

Looking at the keyboard input routine I was looking at understanding where was the conversion of PETSII code to actual screen code. Those line are actually undocumented. I think it would be profitable to add some comments to

.,E733 90 10 BCC $E745 branch if < [SPACE] Nothing to display
.,E735 C9 60 CMP #$60
.,E737 90 04 BCC $E73D branch if $20 to $5F
.,E739 29 DF AND #$DF Conversion of PETCII Character over $60 to Screen code
.,E73B D0 02 BNE $E73F
.,E73D 29 3F AND #$3F Conversion of PETSCII Character from $20 to $5F

Table rows vertically misaligned on Firefox

When viewing the Kernal API, ROM Disassembly, or Memory Map pages with Firefox, the Disassembly/Address row is about half a line further down than the commentary row that it corresponds to. This does not appear to happen in Chrome or IE. See the difference in the attached screenshots:

Mozilla/5.0 (Windows NT 10.0; Win64; x64; rv:83.0) Gecko/20100101 Firefox/83.0:

Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/87.0.4280.88 Safari/537.36:

Transcription around E46F – E4AB can't be right

Something has to be wrong in this section: E46F + 8 = E477, not E473.

Looking at a working ROM, I think the E46F line should only be four bytes long. The four bytes that it quotes after them are jumping ahead to the E473 line.

.:E45F 00 20 42 41 53 49 43 20  basic bytes free
.:E467 42 59 54 45 53 20 46 52
.:E46F 45 45 0D 00 93 0D 20 20
.:E473 93 0D 20 20 20 20 2A 2A  (clr) **** commodore 64 basic v2 ****
.:E47B 2A 2A 20 43 4F 4D 4D 4F  (cr) (cr) 64k ram system
.:E483 44 4F 52 45 20 36 34 20
.:E48B 42 41 53 49 43 20 56 32
.:E493 20 2A 2A 2A 2A 0D 0D 20
.:E49B 36 34 4B 20 52 41 4D 20
.:E4A3 53 59 53 54 45 4D 20 20
.:E4AB 00

A few small errors in c64disasm_en.txt

Recently noticed a few small errors in c64disasm_en.txt. First, at A8E3:

.,A8E3 A2 11    LDX #$02

This instruction should be LDX #$11. Next, at E4CF:

.:E4CF AA AA AA AA AA

There should be only 4 bytes on this line, not 5 (the next instruction is at E4D3). And at EA14:

.,EA14 A9 02    LDA #$02set the count to $02, usually $14 ??

The comment needs nudging over to the 33rd column.

I also noticed that the instructions at E4DA and E621 are LDA $D021 and JSR $E6ED, whereas in a disassembly I've been working on, I have LDA $0286 and JSR $E591. However, I understand that these could just be mistakes in the memory dumps I have, or a difference in ROM versions.

Is a CBM/PET version planned?

Are there plans for a CBM/PET version?

It has its own challenges (BASIC v1, v2 and v4, 40 and 80 columns, etc.),
but there are some good sources available (e.g. Programming the PET/CBM by Raeto West, The Pet Revelead by Nick Hampshire, PET/CBM Personal Computer Guide by Adam Osborne).

License?

Hey, would you mind adding a LICENSE so we know how the content can be used?

I'm planning on creating a web site for aggregating some retro cheatsheets and some of this content would be great for it.

Missing [$nn],Y Addressing Mode

The WDC 65C816 CPU instruction reference chart at https://www.pagetable.com/c64ref/6502/?cpu=65c816&tab=2 appears to be missing all of the Direct Page Indirect Long Y-Indexed addressing mode opcodes.

A13C-A13F is the STEP keyword, rather than STOP

All of the commentaries list the addresses A13C-A13F as the text of the "stop" keyword.

The STOP keyword text is stored at A0DF-A0E2.

The A13C-A13F is the text of the STEP keyword instead.

The same mislabel appears in the German, Lee Davison, and Bob Sander-Cederlof commentary.

Missing root index.html

The umbrella generation script should generate this index.html at the root:
<meta http-equiv="refresh" content="0; URL=/c64ref/c64disasm/" />
...or maybe forward to a different tab.

SID frequencies

I compiled this one a while ago - would you like to adopt it?

https://gist.github.com/matozoid/18cddcbc9cfade3c455bc6230e1f6da6

cleanup the generate code

currently builds in the Sources folder
and is a little all over the place

update readme and contributors while at it?

Scrollbars on Memory Map

Hi, currently you have to scroll all the way to the bottom of the page in order to scroll left and right, because the scrollbar is attached to a div instead of the body.

Removing the overflow: scroll; property line (internal css, line 81) solves the issue.

div.disassembly_container_no_dec {
padding: 1em 0em 1em 11.3em;
overflow: scroll;
}

There is another in div.disassembly_container_with_dec. (line 76).

Hope this helps!