Fireship
May 7, 2026
Operating System Concepts Explained: From Boot to Shutdown
Key Takeaways
- 1Operating systems manage hardware resources through privilege rings, allowing kernels full control while restricting user applications to prevent system crashes
- 2Virtual memory creates the illusion that programs have unlimited memory by translating fake virtual addresses to real physical addresses via the MMU and page tables
- 3Interrupts enable instant responsiveness by allowing hardware to signal the CPU rather than the OS polling continuously for changes
- 4The scheduler manages hundreds of processes on limited CPU cores by fairly distributing CPU time using algorithms like earliest eligible virtual deadline first
- 5System calls are the secure boundary between user applications and kernel operations, requiring privilege escalation from ring 3 to ring 0
- 6File systems organize disk blocks into logical files through index nodes and directories, with journaling preventing corruption from power loss
- 7Interprocess communication mechanisms like pipes allow separate applications to share data safely without shared memory
Chapters
1. Stage 1: The Boot Loader
When the power button is pressed, firmware (UEFI or BIOS) wakes minimal hardware, locates the boot loader (Grub on Linux, IBO on Mac, Bootmgr on Windows), which loads the kernel into RAM and hands off control with full hardware privileges.
2. Stage 2: Privilege Rings
CPUs enforce multiple privilege levels (rings 0-3 on x86). Ring 0 allows kernels unrestricted hardware access, while ring 3 user space requires permission for sensitive operations, protecting programs from interfering with each other.
3. Stage 3: Virtual Memory
The MMU translates fake virtual addresses to real physical addresses using page tables. Each process gets its own page table, creating isolated memory spaces. The TLB caches recent translations, and page faults trigger disk access when needed memory isn't in RAM.
4. Stage 4: File Systems
File systems abstract disk blocks into organized files and folders. Index nodes store metadata and pointers to data blocks but not filenames; directories map filenames to inode numbers. Journaling writes intentions before data, preventing corruption during unexpected shutdowns.
5. Stage 5: Device Drivers and Interrupts
Drivers translate kernel requests into hardware-specific commands for GPUs, Wi-Fi, keyboards, etc. Interrupts are electrical signals that preempt CPU execution, allowing instant response to hardware events without continuous polling.
6. Stage 6: PID1 and Process Creation
The kernel creates PID1 (systemd on Linux), the first user space process and ancestor of all others. Process creation allocates memory, loads executables, sets up virtual address spaces, and adds entries to the process table.
7. Stage 7: System Calls
System calls are the secure API between user applications and kernel operations. They require privilege escalation from ring 3 to ring 0. Linux has ~400 system calls; critical ones include fork and exec for process creation.
8. Stage 8: Process Scheduling
The scheduler manages multiple processes on limited CPU cores by fairly distributing CPU time. Modern Linux uses earliest eligible virtual deadline first (EEVDF) algorithm to ensure each process gets fair access.
9. Stage 9: Threads
Threads share memory and file descriptors but have separate stacks and program counters, allowing single programs to do multiple things simultaneously. Shared memory introduces race condition risks; languages like Go and Rust provide safeguards.
10. Stage 10: Interprocess Communication and Shutdown
Pipes, sockets, and message queues allow safe communication between separate processes. On shutdown, SIGTERM requests graceful process termination, followed by SIGKILL if needed, then file system flushing, driver cleanup, and CPU halt.
Glossary
- UEFI
- Unified Extensible Firmware Interface; modern firmware that initializes hardware before the boot loader.
- BIOS
- Basic Input/Output System; older firmware standard for hardware initialization.
- Bootloader
- Software (like Grub, IBO, or Bootmgr) that finds and loads the kernel into RAM.
- Privilege Rings
- CPU-enforced permission levels (0-3 on x86) that restrict what code can execute; ring 0 is kernel, ring 3 is user space.
- MMU
- Memory Management Unit; hardware that translates virtual addresses to physical addresses.
- Page Table
- Data structure maintained by the kernel that maps virtual addresses to physical memory pages for each process.
- TLB
- Translation Lookaside Buffer; CPU cache that stores recent virtual-to-physical address translations.
- Index Node (Inode)
- File system data structure containing metadata (size, permissions, timestamps) and pointers to data blocks, but not the filename.
- System Call
- API through which user programs request kernel services; triggers privilege escalation from ring 3 to ring 0.
- Process ID (PID)
- Unique identifier assigned by the kernel to each running process for tracking and management.