Awesome

Conker's Bad Fur Day Decompilation

A WIP decompilation of Conker's Bad Fur Day.

Note: To use this repository, you must already have a copy of the game.

Building

Place the US Conker's Bad Fur Day ROM in the root of this repository, name it baserom.us.z64.

Preamble

The assumption is that you will be using Docker for the building process. If this is not the case, see the Dockerfile for the prerequisites; the steps below work perfectly well in Ubuntu 20.04 running via WSL on Windows.

See the Quickstart for more information.

Clone repository

git clone https://github.com/mkst/conker.git --recursive
cd conker

Build Docker image (optional)

docker build . -t conker

Spin up the image interactively (optional)

docker run --rm -ti -v $(pwd):/conker conker bash

Sanity check ROM checksum

make check

Extract ROM

make extract

Decompress code (optional)

make -C conker extract

Compile code (optional)

make -C conker --jobs

Replace compiled code (optional)

make -C conker replace

Compile ROM

make --jobs

If everything matches, you will be greeted with an OK:

build/conker.us.z64: OK

Progress

This project is in its infancy; there are multiple tasks being worked on:

Converting disassembly into (byte-perfect) C code
Extracting assets from the ROM and being able to successfully re-pack them
Identify and document all asset types used in the ROM
Tooling to support the above tasks

Open issues

Identifying and documenting Conker asset (model/texture/sound) format

ROM layout

The layout of the ROM is still a work-in-progress. There are a number of sections within the ROM that are compressed with gzip but have the standard header/trailer stripped and, instead, replaced with a 4-byte header containing the uncompressed data length. These sections are dubbed rzip.

Overview of US ROM shown below:

[header]  0000 0000 > 0000 0040 ; suggests libultra 2.0G
[ boot ]  0000 0040 > 0000 1000 ;
[ code ]  0000 1000 > 0004 2C50 ; init + libultra .text
[ ???? ]  0002 90D0 > ???? ???? ;
[ data ]  0002 C750 > 0002 C7A0 ; init + libultra .data
[ ???? ]  0002 C7A0 > 0004 2C50 ; μcode
[ rzip ]  0004 2C50 > 0018 6B50 ; game .text (compressed)
[ rzip ]  0018 8328 > 0019 C7D8 ; game .data (compressed)
[ code ]  0019 EA88 > 001A 2190 ; debugger .text
[ data ]  001A 2190 > 001A 37E0 ; debugger .data
[ rzip ]  001A 37E0 > 00AB 1950 ; compressed section (textures?)
[ offs ]  00AB 1950 > 00AB 1A40 ; table of asset offsets
[ rzip ]  00AB 1A40 > 03F8 B800 ; assets 00 thru assets 1C
[ ffff ]  03F8 B800 > 0400 0000 ; 0xff padding

Compressed section(s)

There are a number of compressed sections within the ROM. The decompression/compression method is understood and generates matching results.

Building ROM

Due to the compressed code sections, all code segments within the ROM are cut from the ROM and combined together, creating a sub-project inside the conker/ directory.

See the README for more information.

Tools

Custom tools

rarezip/rareunzip; python script to compress/decompress the compression format used in the ROM.

NOTE: gzip is used for compression rather than zlib; use the binary in tools/ in order to get matching compression.

Existing tools

This repo makes use of the following open-source tools without which, there would be no decomp:

asm-differ; compare assembly against the original ROM
asm-processor; allow GLOBAL_ASM wrappers to include assembly within the c files
n64splat; split up the rom & much more...
ido-static-recomp; IDO compiler
gzip; gzip; specifically with the pre-1.5 memzero behaviour

Contributing

The wiki will eventually contain discoveries as they are made.

In the meantime, if you wish to contribute in any way, get stuck in and raise a PR or find me on Discord mkst#4741.