Dinoxor

Re-implementing bitwise operations as abstractions in arm64 neon registers.

Why?

🥴 I was somehow bored and stressed at the same time.

What Is This?

This project demonstrates the use of inline assembly in Rust to implement and benchmark cryptographic algorithms, with a focus on confusing reverse engineers. It includes:

• A dinoxor module for obfuscated XOR operations using inline ARM64 assembly and NEON registers.
• A chacha20 module for implementing the ChaCha20 stream cipher.
• Benchmarking of performance differences between plaintext XOR and inline assembly-based "dinoxor".

Key Features

Inline Assembly in Rust

The project showcases how to:

1. Use the asm! macro for inline assembly
2. Access and manipulate registers (e.g., v0, w0)
3. Call helper functions from within assembly
4. Manage input/output variables and register bindings
5. Manipulate NEON registers for personal gain

ChaCha20 Implementation

• Implements the ChaCha20 stream cipher using Rust's quickcheck for testing.
• Demonstrates how to replace insecure algorithms like RC4 with modern ciphers.

Benchmarks

Performs performance comparisons:

• dinoxor(x, y) vs regular eor x ^ y.
• ChaCha20::process_with_dinoxor vs standard ChaCha20 implementation.

How to Use

Add as a Crate

Add this to your Cargo.toml:

[dependencies]
dinoxor = "*"

Example Usage

use dinoxor::dinoxor::dinoxor;

let result = dinoxor(0b11101011, 0b11111111);
assert_eq!(result, 0b10100);

Benchmarks (from cargo bench)

• Regular XOR: ~6.29 ns/iter
• Dinoxor: ~7,095.95 ns/iter

See Also

• Dinoxor Blog Post
• ChaCha20 Implementation

Developer Notes

• This project is meant to demonstrate how inline assembly can be used for security through obscurity.
• The implementation of ChaCha20 is for educational purposes only.

License

This project is released under the MIT License.