Introduction to Operating Systems¶

What is an Operating System?¶

An operating system (OS) is the software layer that sits between your programs and the hardware.

flowchart TD
    A["User Programs\n(your code)"]
    B["Operating System\n(kernel)"]
    C["Hardware\n(CPU, RAM, Disk, Devices)"]

    A -->|"System calls"| B
    B -->|"Controls"| C
    C -->|"Interrupts"| B

    style A fill:#7c4dff,color:#fff
    style B fill:#1565c0,color:#fff
    style C fill:#37474f,color:#fff

Two Core Jobs¶

Intermediary/Environment Provider: It provides an environment where you can execute programs conveniently and efficiently
Resource Manager: It allocates and manages scarce resources (CPU time, memory, devices, files) so multiple programs/users can safely share the machine

Two-Lens View¶

Extended Machine¶

OS gives you nicer abstractions than raw hardware: - Processes (instead of raw CPU) - Address spaces (instead of physical memory) - Files (instead of disk blocks)

Resource Manager¶

OS coordinates and protects the hardware so the system runs efficiently and correctly.

System Calls¶

How does software cross the boundary into the OS?

Through System calls (open, read, write, fork) which are the "official API" into Kernel services.

To keep things safe, CPUs support at least user mode and kernel mode, switching into kernel mode on a system call/trap/interrupt.

sequenceDiagram
    participant U as User Program
    participant K as Kernel

    U->>K: System call (e.g. read())
    Note over K: Switch to kernel mode
    K->>K: Execute privileged operation
    K->>U: Return result
    Note over U: Switch back to user mode

Kitchen Analogy¶

The CPU is the stove, RAM is counter space, disks are pantry storage, and devices are appliances.

The OS is the head chef and kitchen manager: - You (a program in user mode) don't walk into the kitchen and touch the stove wiring directly - You place an "order" via a system call - The manager decides when you get stove time, how much counter space you get, and stops you from wrecking someone else's meal

This is protection via user/kernel mode.

Processes¶

A process is a program in execution plus everything the OS needs to run it: CPU state, memory mapping, open files, etc.

The OS tracks each process using a Process Control Block (PCB) which stores at minimum: - Process State - Program counter - CPU registers - Scheduling info - Memory-management info - I/O Status like open files

Process States¶

stateDiagram-v2
    [*] --> New
    New --> Ready : admitted
    Ready --> Running : scheduler dispatch
    Running --> Ready : interrupt
    Running --> Waiting : I/O or event wait
    Waiting --> Ready : I/O or event complete
    Running --> Terminated : exit
    Terminated --> [*]

Only one process can be running per CPU core at a time. Others are ready or waiting.

Context Switch¶

A context switch happens when the OS stops one running process and resumes another by saving/restoring its CPU state (registers + PC) in/from the PCB.

Interrupts (especially timer interrupts) are a classic trigger that forces the OS to regain control and potentially switch the running process.

Observing Process State¶

# 1. Start a CPU hog
yes > /dev/null &

# 2. Get its PID
echo $!

# 3. Inspect it with ps
ps -p <PID> -O pid,ppid,stat,comm

# 4. Watch it live
top -p <PID>

# 5. Stop/continue/kill it
kill -STOP <PID>
kill -CONT <PID>
kill <PID>

Stopped vs Sleeping/Waiting¶

Stopped (SIGSTOP) - The process is paused on purpose by a signal - It could run but the OS says "you're not allowed to run right now" - Will stay paused until it gets SIGCONT or you kill it

Sleeping/Waiting - The process literally cannot make progress yet - Waiting for something external: keyboard input, disk read, network packet, or a lock

Important for Scheduling¶

Stopped: scheduler ignores it because it's forbidden to run
Waiting: scheduler ignores it because it's unable to run