Overview of Operating Systems

What’s the point of the syscall interface in a monolithic kernel, if a monolithic kernel is always in kernel mode i.e. users directly interact?
→ could be hole in understanding of what a “syscall” is; previous understanding is that a syscall has a transition between user and kernel mode because safety
Previous understanding from “When a user process requires a service from the kernel (e.g., reading from a file or writing to a file) it invokes a system call”

• MS-DOS is an old OS by Microsoft

Operating System Structures

Monolithic Kernels

Layered Approach

Advantages:

• Simplicity of construction
• Ease of debugging (and therefore safety)
• Clear interfaces between layers
  • ease of use

Disadvantages:

• Defining layers (separation) is difficult
• Efficiency: adds a bit of overhead to go between the layers

Microkernels

Mach OS is one of the earliest examples of a microkernel.
As the name suggests:

• Removes all non-essential components from the kernel → smaller kernel
• contrasts against monolithic kernels, which are generally MASSIVE (do everything)

Advantages:

• Extending from the OS is easy
  • We’ll see how you can use Loadable Kernel Modules
• Safe and reliable

Disadvantages:

• Performance overhead (syscall overhead)