Free chapter — no signup required
Operating Systems, Kernel & Hardware Sympathy
Hardware Sympathy
Introduction: Thinking Like a Mechanic
The Great Lie: Memory is Fast
The Latency Hierarchy (The Desk Analogy)
Cache Lines and Spatial Locality
False Sharing (The Noisy Neighbor Problem)
Branch Prediction and Speculation
Spectre and Meltdown: When Speculation Becomes a Security Hole
NUMA (Non-Uniform Memory Access)
Putting It All Together: Hardware Sympathy in Practice
The Linux Filesystem
Linux Permissions Model
Introduction: The Security Gatekeeper
The "rwx" Triad: Read, Write, Execute
Viewing Permissions: The `ls -l` Command
Changing Permissions: `chmod` (Symbolic Mode)
Changing Permissions: `chmod` (Octal Mode)
Changing Ownership: `chown` and `chgrp`
Default Permissions: `umask`
Special Permissions: The Sticky Bit
Special Permissions: SUID (Set User ID)
Special Permissions: SGID (Set Group ID)
Combining Special Bits: The Four-Digit Octal
Access Control Lists (ACLs): Beyond Basic Permissions
Process Internals: Lifecycle, States & Signals
Introduction: What is a Process Really?
The Lifecycle Part 1: Birth by Cloning (`fork`)
The Lifecycle Part 2: Transformation (`exec`)
The Lifecycle Part 3: Death and Reporting (`exit` & `wait`)
Process States: Running and Sleeping
Process States: The Zombie (Z)
Process States: The Deep Sleep (D-state)
Understanding Load Average
Signals: Communicating with Processes
The "Polite" Signals: SIGTERM and SIGHUP
The Nuclear Option: SIGKILL
Systemd: The Linux Service Manager
Introduction: The Orchestra Conductor
Unit Files: The Instructions Systemd Reads
Best Practice: Overrides and Drop-In Files
Types of Units: Services, Timers, Mounts, and More
Dependency Management: Wants, Requires, and After
The Nuclear Option: Masking Services
Targets: Grouping Units Together
The Journal: Centralized Logging with journalctl
The Linux Boot Process
Resource Management & Isolation
Linux Memory Management
Introduction: The Grand Illusion
Paging and Page Tables: The Map
Page Faults: When the Map Fails
Memory Overcommitment: The Calculated Gamble
Swap: The Overflow Parking Lot
Swappiness: Controlling the Trigger
The OOM Killer: The Grim Reaper
Influencing the OOM Killer
Memory Visibility Tools: Reading the System's Mind
Linux Kernel Scheduling
Introduction: The CPU's Traffic Cop
Scheduling Classes: A Hierarchy of Priorities
The Default: Completely Fair Scheduler (CFS)
How CFS Works: Virtual Runtime (vruntime)
Context Switching: The Cost of Multitasking
Preemption Models
Real-Time Scheduling: Cutting the Line
RT Policies: SCHED_FIFO vs. SCHED_RR vs. SCHED_DEADLINE
CPU Affinity and NUMA Awareness
Linux Observability & Tracing
Introduction: Seeing Inside the Black Box
The Old Guard: System Call Tracing with `strace`
The Network Sniffer: `tcpdump`
The Revolution: Introduction to eBPF
eBPF Architecture & Safety: The Verifier
Modern eBPF Tools I: BCC (BPF Compiler Collection)
Modern eBPF Tools II: bpftrace
The Broader Observability Ecosystem: perf and Flame Graphs
The Storage Stack & I/O Path
Introduction: The Journey of a Write
The Virtual Filesystem (VFS): One API to Rule Them All
Buffered I/O vs. Direct I/O
The Block Layer and the `bio` Struct
I/O Schedulers: The Elevator Algorithm
Synchronous vs. Asynchronous I/O
Generation 1: Select and Poll
Generation 2: Epoll (The Event List)
Generation 3: io_uring (The New King)
Kernel Networking Scaling
Inter-Process Communication (IPC)
Timekeeping in Linux
Security Modules (LSM)
Virtualization & Cloud Primitives
System Reliability & Tuning
Advanced Inspection Tools
Binary Anatomy & Linking
Interrupt Handling
Filesystem Integrity
CLI Literacy
Windows Server Basics
Cloud-Init & User Data
GPU/TPU Kernel Interactions
Introduction: The Computer Inside Your Computer
The Driver Stack: Kernel vs. User Space
PCIe Topology and NUMA Affinity
Unified Memory and GPU Page Faults
NVLink vs. PCIe for Multi-GPU Communication
GPUDirect RDMA — Bypassing the CPU Entirely
Debugging with nvidia-smi — Basic Monitoring
Debugging with nvidia-smi — Topology and Affinity
Xid Errors — When Hardware Fails
Pressure Stall Information (PSI)
SystemTap
HugePages: Transparent & Explicit
This is one of 18 chapters
Every chapter is free — Kubernetes, Terraform, SRE, distributed systems, and more — with fast daily review built in.
Sign up free