A mini x86-64 assembler for fun

Last update: Dec 9, 2022

Related tags

Overview

minias

A mini assembler for x86-64, written for fun and learning.

Minias can assemble itself and many/most things compiled with the cproc C compiler i.e., large amounts of real world software.

Project Goals:

A simple, tiny, fast implementation (in that order).
Assemble the output of cproc/qbe and chibicc.
Relocatable elf output.

Non Goals:

Assemble every assembly instruction.
Assemble other architectures.
Work as a library.

Building

Install the peg/leg parser generator, make and a C compiler and run:

make

leg asm.peg > asm.peg.inc
cc -O2 *.c -o minias

Roadmap

Essential features:

Self host with cproc.
Self host with chibicc.

Bonus features:

Two pass jump relaxing.
Immediate relaxing.
Simple immediate expressions.
Assemble a libc.
Test every opcode with all variants in our test suite.
Parser that doesn't depend on peg/leg.

Notes

Minias deliberately does not free allocated memory as it all is freed by the OS at the end of execution. Memory usage is still quite light as it uses string and value interning. In the future we could use an arena allocator for minias and still avoid manual calls to free.
Minias deliberately keeps the peg grammar quite repetitive and simple, please keep it this way.
Performance is limited by the parser, it would be interesting to see if we can improve the parser generator upstream. That being said, performance is often better than gnu as and much better than the clang assembler,
One day it would be nice to write a 'minipeg' in a single .c file that can be bundled in projects.

Contact

Ask questions on the mailing list. Submit bugs to the ticket system.

Resources

intel reference - Specifically chapter 2.1 and chapter 3.1.
elf spec
osdev wiki
goas
neatas

funky fun fuzzer

FunkyFunFuzzer / fffz FunkyFunFuzzer / fffz is an attempt at a file fuzzer prioritising usability while striving for real-world performance. It is a m

11 Oct 8, 2021

A basic example of IoT for fun

esp32 simple led IoT A basic example of IoT for fun about Im using platformio for upload this project so u should be used that if you want to use this

4 Mar 9, 2022

A fast and fun layout manager for FLTK

Fl_Flow A fast and fun layout manager for FLTK Fl_Flow is a layout manager for FLTK which works in a largely different manner to existing systems you

40 Dec 1, 2022

Having some fun with the C++ programming language.

Having some fun with the C++ programming language. Coding Environment Text editor. Visual Studio Code for Windows Compiler. MingW GCC 6.3.0 Coding sty

4 Jan 1, 2023

Just a password tool that I make for fun and the sake of my fascination

Note I made this in 1 day and I only have 1 years of experience, feel free to critique my spaghetti of a code. And if 15 chars of password isn't enoug

2 Dec 11, 2021

MasterPlan is a project management software / visual idea board software. It attempts to be easy to use, lightweight, and fun.

MasterPlan is a customizeable graphical project management software for independent users or small teams. If you need to share plans across a whole co

444 Dec 23, 2022

Comments

Use smaller encodings in a few more cases.

None of these are bugs, but they showed up when comparing disassembly between gas and minias.

I'm less confident about the imul change than the other two. Please double-check that it looks correct to you.

opened by michaelforney 0
Use void * for functions that access bytes of object representation.

Although out() and secaddbytes() only use their argument to call fwrite()/memcpy(), using the type uint8_t * suggests that it might be dereferenced as uint8_t. In general, it is undefined behavior to access an object through a uint8_t pointer, since uint8_t is not necessarily a character type.

So, just use void * to avoid some casts and implication of possible undefined behavior.

opened by michaelforney 0
Use Elf64_Rela for relocations.

I needed this to run hello world on my system (otherwise I get a segfault). I'm not sure exactly why Elf64_Rel worked for you but not for me, but since the spec says to always use explicit addends, this seems like the right thing to do.

opened by michaelforney 0