Skip to content

README.md

Latest commit

 

History

History
282 lines (217 loc) · 9.06 KB

README.md

File metadata and controls

282 lines (217 loc) · 9.06 KB

🧬 Chapter 18: Kernel & System Internals

Advanced Chapter 18

📑 Table of Contents

Kernel Architecture

The Linux kernel is monolithic — all core services run in kernel space as a single binary.

┌────────────────────────────────────────────────┐
│                USER SPACE                       │
│  Applications → Libraries → System Call API     │
├────────────────────────────────────────────────┤
│                KERNEL SPACE                     │
│  ┌──────────────────────────────────────────┐  │
│  │ System Call Interface                     │  │
│  ├────────┬────────┬────────┬───────────────┤  │
│  │ VFS    │Process │ Memory │ Network       │  │
│  │(Virtual│ Sched  │ Mgmt   │ Stack         │  │
│  │ FS)    │        │        │               │  │
│  ├────────┴────────┴────────┴───────────────┤  │
│  │         Device Drivers (modules)          │  │
│  ├──────────────────────────────────────────┤  │
│  │  Architecture-Dependent Code (x86, ARM)   │  │
│  └──────────────────────────────────────────┘  │
├────────────────────────────────────────────────┤
│                HARDWARE                         │
└────────────────────────────────────────────────┘

Kernel Subsystems

Subsystem Role
Process Scheduler Decides which process runs on which CPU core
Memory Manager Virtual memory, paging, OOM killer
VFS Abstraction layer over all filesystems
Network Stack TCP/IP, sockets, netfilter
Device Drivers Hardware communication (loaded as modules)
IPC Inter-process communication (pipes, signals, sockets) 
# Kernel information
uname -r                        # Kernel version (e.g., 6.8.0-41-generic)
uname -a                        # All system info
cat /proc/version                # Kernel build details
cat /proc/cmdline                # Boot parameters

The Boot Process

1. BIOS/UEFI
   └→ POST (Power-On Self-Test)
   └→ Find boot device

2. Bootloader (GRUB)
   └→ Load kernel + initramfs into memory
   └→ Pass boot parameters to kernel

3. Kernel
   └→ Decompress itself
   └→ Initialize hardware, memory, CPU
   └→ Mount initramfs as temporary root
   └→ Execute /init from initramfs

4. initramfs
   └→ Load necessary drivers (disk, filesystem)
   └→ Find and mount real root filesystem
   └→ pivot_root to real filesystem

5. Init System (systemd, PID 1)
   └→ Mount filesystems from /etc/fstab
   └→ Start services (networking, SSH, etc.)
   └→ Reach default target (multi-user or graphical)
   └→ Display login prompt

GRUB Bootloader

# GRUB config
cat /boot/grub/grub.cfg              # Generated config (don't edit!)
cat /etc/default/grub                # User settings

# Edit GRUB settings
sudo vim /etc/default/grub
# GRUB_TIMEOUT=5
# GRUB_CMDLINE_LINUX_DEFAULT="quiet splash"
# GRUB_CMDLINE_LINUX=""

# Rebuild GRUB config
sudo update-grub                      # Debian/Ubuntu
sudo grub2-mkconfig -o /boot/grub2/grub.cfg  # RHEL/Fedora

# Reinstall GRUB (boot repair)
sudo grub-install /dev/sda
sudo update-grub

Kernel Modules

Modules are pieces of kernel code loaded on demand (like drivers).

# List loaded modules
lsmod
lsmod | grep usb

# Module info
modinfo ext4
modinfo nvidia

# Load a module
sudo modprobe vfat               # Load FAT filesystem module
sudo modprobe -v snd-hda-intel   # Verbose load

# Unload a module
sudo modprobe -r vfat            # Remove module
sudo rmmod vfat                  # Alternative

# Load module at boot — add to:
echo "vfat" | sudo tee /etc/modules-load.d/vfat.conf

# Blacklist a module (prevent loading)
echo "blacklist nouveau" | sudo tee /etc/modprobe.d/blacklist-nouveau.conf
sudo update-initramfs -u         # Rebuild initramfs

# Module parameters
modinfo i915 | grep parm         # Show available parameters
echo "options i915 enable_guc=3" | sudo tee /etc/modprobe.d/i915.conf

# Module dependencies
modprobe --show-depends ext4

sysctl — Runtime Kernel Parameters

# View all parameters
sysctl -a

# View specific parameter
sysctl net.ipv4.ip_forward
sysctl vm.swappiness

# Set temporarily (until reboot)
sudo sysctl -w net.ipv4.ip_forward=1
sudo sysctl -w vm.swappiness=10

# Set permanently
echo "net.ipv4.ip_forward = 1" | sudo tee -a /etc/sysctl.conf
sudo sysctl -p                   # Reload

# Or use drop-in files
echo "vm.swappiness = 10" | sudo tee /etc/sysctl.d/99-custom.conf
sudo sysctl --system              # Reload all

Common sysctl Tweaks

Parameter Description Default
vm.swappiness How aggressively to use swap 60
net.ipv4.ip_forward Enable IP forwarding (routing) 0
net.core.somaxconn Max socket connections queue 4096
fs.file-max Max open files system-wide 9223372036854775807
fs.inotify.max_user_watches Max inotify watches 8192

initramfs / initrd

The initramfs is a temporary root filesystem loaded into RAM during boot, before the real root is mounted.

# View current initramfs
ls -la /boot/initrd.img*

# List contents
lsinitramfs /boot/initrd.img-$(uname -r)

# Rebuild initramfs
sudo update-initramfs -u              # Update current kernel
sudo update-initramfs -u -k all       # Update all kernels
sudo mkinitramfs -o /boot/initrd.img-$(uname -r)  # Manual rebuild

# Dracut (RHEL/Fedora)
sudo dracut --force

System Calls

System calls are the API between user space and kernel space.

# Trace system calls of a command
strace ls /tmp                   # See all syscalls ls makes
strace -c ls /tmp                # Summary statistics
strace -e open,read,write ls     # Filter specific syscalls
strace -p 1234                   # Attach to running process
strace -f -e trace=network curl https://example.com  # Network calls

# Common system calls
# open()   — Open a file
# read()   — Read from file descriptor
# write()  — Write to file descriptor
# close()  — Close file descriptor
# fork()   — Create child process
# exec()   — Execute a program
# mmap()   — Map memory
# stat()   — Get file info
# socket() — Create network socket

The init System (PID 1)

# Check your init system
ps -p 1 -o comm=                 # systemd, init, or others
stat /proc/1/exe                 # Binary path

# systemd (modern standard)
systemctl list-units              # All loaded units
systemctl list-unit-files         # All available units
systemctl get-default             # Default boot target
systemctl set-default multi-user.target   # Boot to CLI
systemctl set-default graphical.target    # Boot to GUI

# Boot targets (runlevels)
# multi-user.target = runlevel 3 (CLI)
# graphical.target  = runlevel 5 (GUI)
# rescue.target     = single user
# emergency.target  = minimal shell

🏋️ Practice Exercises

  1. Kernel: Check your kernel version and when it was compiled
  2. Modules: List all loaded modules and find one related to your network
  3. modinfo: Get detailed info about the ext4 module
  4. sysctl: Check and change the swappiness value
  5. strace: Trace the system calls of cat /etc/hostname
  6. Boot: Read /proc/cmdline to see your boot parameters
  7. GRUB: Look at /etc/default/grub settings
  8. initramfs: List the contents of your current initramfs

← Previous: Advanced Shell Scripting · 🏠 Home · Next: Systemd & Service Management →