MACROS

The syscall() macro should be invoked as follows:
	syscall(#,arg1,arg2,arg3,arg4,arg5,arg6)

Error checking is built into the macro. You _MUST_ define an error-handler.
If the syscall() macro is invoked without an error handler defined, the
assembler will throw an error.

You define error handlers like this:

	ech(label_for_your_error_code)
	; do shit
	ech(other_label_for_stuff)

The error checking consists of:

	or rax,rax
	js @ERROR_HANDLER

So, the FLAGS are set based on the error return of the syscall, which is nice.

The arguments are loaded with 32-bit 'mov' by default.
You can change that in the following ways:
	+arg	: Load arg with a 64-bit 'mov'
	^arg	: Load arg with a 32-bit 'mov' (redundant in this usage)
	^^arg	: Load arg with a 16-bit 'mov'
	^^^arg	: Load arg with an 8-bit 'mov'
	[arg]	: Load arg with 'lea' (eg, "lea rdi,[arg]")
	+[arg]	: Dereference arg and load (eg, "mov rdi,[arg]")
	^[arg]
	^^[arg]
	^^^[arg]
	[+arg]	: Discard brackets and load with 'mov' (eg, "mov rdi,arg")
	[^arg]	  (This functionality is stupid)
	[^^arg]
	[^^^arg]
	NULL	: Make the argument NULL (Always 64-bit, eg "xor rdi,rdi")

Note that the syscall number ("#" in the example) is also expanded in this fashion

--

In our thread-based model, each thread calls mmap() to grab it's own memory,
and sets %rbp the base of that memory.

The tmalloc(symbol,$bytes) macro "reserves" space in that memory. It defines
'symbol' to be [rbp+bytes].

So, for example:

tmalloc(crap,8)
tmalloc(shit,16)
tmalloc(moreshit,32)

crap is "rbp"
shit is "rbp+8"
moreshit is "rbp+24"