Operating Systems¶

Operating system theory and concepts for computer science studies

Scope and use¶

This index links to teaching models, not configuration advice for a specific kernel. Algorithm properties assume the workload stated in each guide; implementation claims and timings require an OS/kernel version, hardware, workload, and trace. A lab is complete only when inputs, assumptions, failure states, commands, and raw results are recorded. Retry failed experiments only after preserving the error and changing one controlled condition.

Topics¶

CPU Scheduling

Process scheduling algorithms: FCFS, SJF, Priority, Round Robin, MLFQ.

View Guide
Synchronization

Mutex, semaphores, monitors, and critical section problems.

View Guide
Deadlocks

Deadlock conditions, prevention, avoidance, and detection.

View Guide
Memory Management

Virtual memory, paging, segmentation, and page replacement.

View Guide
Process Management

Process creation, fork(), exec(), and inter-process communication.

View Guide
Virtualization

Hypervisor mechanisms and Docker vs VM isolation comparison.

View Guide
Distributed Deadlocks

Deadlock handling in distributed systems and Edge Chasing algorithm.

View Guide

Learning Path¶

flowchart TD
    subgraph Fundamentals
        A[Process Concept] --> B[Process States]
        B --> C[Context Switching]
    end

    subgraph Scheduling
        C --> D[CPU Scheduling]
        D --> E[Scheduling Algorithms]
    end

    subgraph Concurrency
        E --> F[Synchronization]
        F --> G[Critical Section]
        G --> H[Deadlocks]
    end

    subgraph Memory
        H --> I[Memory Management]
        I --> J[Virtual Memory]
        J --> K[Page Replacement]
    end

Key Concepts¶

Process States¶

stateDiagram-v2
    [*] --> New: Create
    New --> Ready: Admit
    Ready --> Running: Dispatch
    Running --> Ready: Interrupt
    Running --> Waiting: I/O Wait
    Waiting --> Ready: I/O Complete
    Running --> Terminated: Exit
    Terminated --> [*]

Scheduling Algorithms Comparison¶

Algorithm	Preemptive	Starvation	Best For
FCFS	No	Possible with an unbounded/unstable arrival model	Simple arrival-ordered workloads
SJF / SRTF	SJF: No; SRTF: Yes	Yes, without aging/admission bounds	Workloads with known or estimated bursts
Priority	Policy-dependent	Yes, without aging or reservation	Explicit service classes; not sufficient alone for real-time guarantees
Round Robin	Yes	Avoided only under finite runnable set, positive quantum, and equal service class	Time-sharing
MLFQ	Yes	Possible without periodic boost or allocation bounds	Adaptive general-purpose scheduling

Deadlock Conditions¶

For the classic reusable, non-preemptible resource model, all four Coffman conditions are necessary for deadlock:

Mutual Exclusion - Resource held exclusively
Hold and Wait - Process holding while waiting
No Preemption - Cannot forcibly remove
Circular Wait - Circular chain of processes

Quick Formulas¶

These formulas use the simplified single-CPU examples in the guides. Waiting time equals turnaround minus CPU burst only when omitted I/O/service intervals and overhead are handled consistently. Address formulas assume integer page/frame numbering and the stated page size.

CPU Scheduling¶

Turnaround Time = Completion Time - Arrival Time
Waiting Time = Turnaround Time - Burst Time
Response Time = First Response - Arrival Time
Throughput = Number of Processes / Total Time
CPU Utilization = (CPU Busy Time / Total Time) × 100%

Memory Management¶

Page Number = Virtual Address / Page Size
Page Offset = Virtual Address % Page Size
Physical Address = (Frame Number × Page Size) + Offset