Self-modifying code is something there doesn't seem to be much talk about, so here's some basic examples of it!
Current examples:
- Basic control flow
- Basic printf
- Factorial
- Hello, World!
There's currently only x86 linux examples currently, I may look into windows and llvm in the future.
Go into the directory of the code you want to run and execute ../execute.sh code-name. For example: ../execute.sh factorial.
- nasm
- ld
- objcopy
- objdump
- libc:i386
If you look at the execute script, you'll notice that it's a bit odd. ld has --omagic to allow the .text section to be written, but that also disables dynamic linking. objcopy is used to make the .text section writable, but because of a bug, another script from stackoverflow is used to fix it.
The normal method of writing self-modifying code is to allocate your own memory and do a dance where sometimes it's executable and sometimes its writable. That's a pain, so the program just tells the system that .text is modifiable through a flag here.
Not all of the code will follow proper calling conventions and have a regular formatting. I'm a reverse engineer, not an assembly programmer.