Correct small bug in memcpy 32bits implementation

I'm not entirely sure about it, but the rest of the file treats 32 bits as groups of 4 so it makes sense that memcpy does the same.

Reduce kernel heap to 64 MB
Fix issue in build.rs

Add temporary override for spin-rs

Add commented out system call debug printing code

Add syscall name debugging
Update debugging code

Improve error information
Cleanup interrupt macros

Fully implemented AML parser, some amendments to ACPI infrastructure

Support arguments in #!

Implement F_DUPFD

Prevent freezing due to double locking