changelog.txt

	- Added a basic ELF parser, along with a hello world user application.

	- Scheduler : The scheduler is now capable of loading and executing user tasks. The
	MMX, FPU and SSE registers are unfortunately not saved/restored yet, so
	applications are (for now!) not allowed to use them.

	- Scheduler : Every task has an assigned kernel stack. This kernel stack is used to
	executes syscalls, IRQs etc. The kernel stack is always mapped to the exact
	same address, and each task has a different physical page mapped into that
	address. This allows for every interrupt to execute on a different stack,
	even though they all use the same address, by simply taking advantage of
	paging.

	- Scheduler : Likewise, the user stack is also allocated and used in the same
	way. I.e. it is always mapped to the same address, and different pages are
	mapped for different tasks. See doc/kernel_stack.txt for more info, although
	it does not go into much detail about actual kernel stacks yet.

	- Scheduler : There is now a task loader function (written in assembly) that
	executes at the start of a task (before actually jumping to the task).
	Its whole purpose is to unlock the scheduler before the task executes, and
	then it jumps.

	- Heap : Fixed minor bug that caused more pages to be mapped than what was
	actually allocated for the heap.

	- Syscalls : There is now a syscall interface. Interrupt number 0x80 is
	reserved for this interface. Currently, only exit(), read(), write(), open()
	close(), fork(), wait() and exec() are implemented. This is good enough for simple shell
	commands, as demonstrated in the sample program that is loaded on start-up.
	(in /bin/sh)
	Hopefully, I can port another, better shell soon (and a libc of course).

	- Task termination : The ability to terminate tasks has been added. There is
	always a new task created at start-up. This task has the sole purpose of
	cleaning up tasks in the termination queue. The terminate_task() function
	is there for processes that want to exit. A syscall (rax=0 rbx=err_code) exit
	is provided for user-space tasks.

	- Standard input-output : the application loaded by the kernel at bootup
	(currently root/bin/sh) has 2 file descriptors open at start-up, 0 referring
	to standard input (the keyboard) and 1 referring to standard output (screen)

	- Keyboard : The way keyboard input is handled has been changed. When a
	keyboard IRQ comes in, the character is first printed to screen and
	added to a buffer (if not enter) and if it is a newline, then the buffer is
	flushed to a vnode. The first loaded application has a file descriptor open to said vnode.