System Calls 
	Userspace programs can request action by the kernel
		- Like with /proc files
		- System calls have return AND parameters	
	CS250 review:  How to trigger one of these
	POSIX:  Portable Operating System Interface X
		- man lists POSIX compatibility
		- *nix systems are POSIX-compatible
		- Older Windows is, maybe still by library (nanosleep)
		- Philisophically, non-busybody, like big box store
	Remember:  On *nix, return of 0 generally means success
		- Otherwise, errno will be set
	Need two pieces:
		- Userspace library function
		- Kernel implementation
	Let's look at read
		- First:  What do we do to call it in the program?
		- read system call (fs/read_write.c)
		- vfs_read and what is vfs (file_operations struct)
		- Macro sets up system call, assigns number, etc.
		- Exact set of system calls is architecture-dependent
		- read in arch/x86 system call table
	Mechanism:
		- Memory manipulation as required (like x86 function call)
		- System call number goes in eax
		- Interrupt 0x80 is triggered (we'll play with interrupts later)
		- We end up in entry_64.S
		- System call is looked up, and the appropriate kernel code is run
	Generally speaking, system calls have a particular purpose
		- Like read
		- NOT like ioctl
	
	The book goes through the process of adding one
		- This class (usually) covers using /proc and /dev, but not syscalls
	
	Why not add these?
		- After all, they're fast, and pretty simple to add actually
		- Syscalls have syscall numbers that are part of the kernel interface
		- If you add a system call, on an "unofficial" number, and
				that number is later used, you will have a problem
		- Can't be used without userspace C code (no vfs interface)
		- /proc, /dev, and /sys exist for this (/sys is newer than me)