The Practice Configuration Table

The table below is our hands-on exercise for this lecture. Each row shows a config option, its location in the menuconfig menu, what we are changing it to, and why. We will cover each one in detail in the sections below. All paths are verified against Linux 6.x.

Option 1: CONFIG_LOCALVERSION — Naming Your Custom Kernel

Every Linux kernel has a version string. When you build a custom kernel — a key practical skill in this free Linux kernel development course — you want to be able to identify your build among other kernels on the system. CONFIG_LOCALVERSION lets you add a custom suffix to that version string.

Navigate to General setup → Local version – append to kernel release. Press Enter on the ( ) field, type -embeddedpathashala, and press Enter to confirm.

# After building and booting your kernel:
$ uname -r
6.12.0-embeddedpathashala

Option 2: CONFIG_IKCONFIG — Embedding the Build Config in the Kernel

Have you ever run a kernel on a device and wondered exactly what configuration options it was built with? CONFIG_IKCONFIG solves this by embedding a compressed copy of the .config file directly inside the kernel image itself.

Once enabled, you can extract the config from any kernel image — even if you no longer have the source tree — using the scripts/extract-ikconfig script:

# Extract config from a kernel image on disk:
scripts/extract-ikconfig /boot/vmlinuz-6.12.0-embeddedpathashala | less

# Or, if CONFIG_IKCONFIG_PROC is also enabled (see next section):
zcat /proc/config.gz | grep CONFIG_UIO

Option 3: CONFIG_IKCONFIG_PROC — Live Config Access via /proc

This is a child option of CONFIG_IKCONFIG. When both are enabled, the kernel’s embedded configuration becomes accessible as a gzip-compressed pseudo-file at /proc/config.gz. You can read it on any running system without needing the original source tree or build directory.

Many Linux distributions (Android, Arch Linux) enable both options so kernel developers can always inspect the exact configuration of any running kernel image.

Option 4: CONFIG_PROFILING — Kernel Performance Profiling

Kernel profiling support enables the infrastructure needed by performance analysis tools like perf and OProfile. It allows you to measure where the kernel spends its CPU time, which code paths are hot, and which functions are bottlenecks.

We disable it in this exercise (n) purely to practice turning off a default-on option. In real development work, keep this enabled — perf is one of the most powerful tools in a Linux kernel developer’s toolkit.

Option 5: CONFIG_HAMRADIO — Amateur Radio Support

The Linux kernel includes support for Amateur (HAM) Radio hardware and the AX.25 packet radio protocol. This is a great example of a broad theme in this free Linux kernel development course: the kernel contains code for an enormous variety of hardware, and most of it will never be used on any given system. For typical x86 development machines or embedded targets, HAM radio hardware simply does not exist.

Disabling this option (n) reduces the kernel’s config surface area with zero practical downside for the vast majority of systems. This is healthy kernel hygiene — include only what you actually need.

Option 6: CONFIG_VBOXGUEST — VirtualBox Guest Integration

If you are learning kernel programming inside a VirtualBox virtual machine — a common setup in this free Linux kernel development course — this driver provides guest integration features: shared clipboard, better mouse integration, and dynamic screen resolution changes.

We set it to m (module). This is the ideal choice: on a VirtualBox guest the module is loaded automatically; on physical hardware it simply does not load. Zero overhead either way, and no need to decide at build time whether you are in a VM or not.

Option 7: CONFIG_UIO — Userspace I/O Framework

UIO (Userspace I/O) is a kernel framework that allows most of a device driver’s logic to run in userspace rather than in kernel space. A thin kernel module handles only interrupt notifications; all device communication, register access, and data processing happens in a normal userspace application. This makes driver development simpler, safer (bugs crash the app, not the kernel), and easier to debug with standard userspace tools.

CONFIG_UIO_PDRV_GENIRQ extends this for platform devices (common on ARM/RISC-V SoCs) that use a standard generic interrupt handler. Enable it to m alongside CONFIG_UIO.

Option 8: CONFIG_MSDOS_FS — MS-DOS Filesystem

The MSDOS filesystem is the original FAT12/FAT16 format without long filename support. It is rarely used on modern primary storage, but still appears on floppy disk images, some embedded bootloader partitions, and very old removable media. We enable it as a module (m) so the kernel can mount such filesystems when needed, without the code occupying memory all the time.

Note: If you need standard FAT32 support for modern USB drives, that is a different option: CONFIG_VFAT_FS. MSDOS FS and VFAT FS are different drivers in the kernel.

⚠️ Option 9: CONFIG_SECURITY — Linux Security Modules

CONFIG_SECURITY enables the Linux Security Modules (LSM) framework — the pluggable architecture that makes mandatory access control systems like SELinux, AppArmor, TOMOYO, and SMACK possible. These go far beyond standard Unix file permissions and are fundamental to modern Linux security.

Option 10: CONFIG_DEBUG_STACK_USAGE — Stack Instrumentation

Every kernel thread has a fixed-size stack — typically 8KB on 32-bit systems and 16KB on 64-bit x86. Stack overflows in kernel space are catastrophic because the kernel has no mechanism to handle them gracefully — they cause silent data corruption or a kernel panic. CONFIG_DEBUG_STACK_USAGE adds lightweight instrumentation to track how much of each thread’s stack is actually consumed, helping developers identify dangerous code paths.

This is essential in any free Linux kernel development course — when you are writing new kernel code, you need to know you are not approaching the stack limit.

Verifying All Your Changes

After making changes in menuconfig and saving, verify the resulting .config file before starting the build. Use grep to check multiple options at once:

# Verify all our changes in one command:
grep -E "CONFIG_LOCALVERSION|CONFIG_IKCONFIG|CONFIG_UIO|CONFIG_VBOXGUEST|\
CONFIG_SECURITY|CONFIG_DEBUG_STACK_USAGE|CONFIG_PROFILING|CONFIG_HAMRADIO" .config

Expected output after completing this lecture’s exercise:

CONFIG_LOCALVERSION="-embeddedpathashala"
CONFIG_IKCONFIG=y
CONFIG_IKCONFIG_PROC=y
# CONFIG_PROFILING is not set
# CONFIG_HAMRADIO is not set
CONFIG_VBOXGUEST=m
CONFIG_UIO=m
CONFIG_UIO_PDRV_GENIRQ=m
# CONFIG_MSDOS_FS is not set
# CONFIG_SECURITY is not set
CONFIG_DEBUG_STACK_USAGE=y

Once you are satisfied, start the kernel build:

# Build using all available CPU cores (fastest):
make -j$(nproc)

# Check how many cores you have:
nproc

Bonus: Other Useful Configuration make Targets

Common Mistakes and Security Considerations

• Disabling CONFIG_SECURITY on anything other than a throwaway test kernel removes all LSM-based security (AppArmor, SELinux). Never do this in production.
• Forgetting make modules_install after building. Your kernel may boot but modules will fail to load if they are not installed to the correct path.
• Mixing module versions — always run make modules_install for the same kernel version you are about to boot. Running mismatched modules causes load failures.
• Not backing up .config before make mrproper. This is the most common way to lose hours of configuration work.
• Enabling too many debug options simultaneously on embedded targets. Options like CONFIG_KASAN significantly increase memory usage — enable them one at a time.

Frequently Asked Questions — Linux Kernel Configuration Options

References and Further Reading

• Linux Kernel Documentation — Kconfig language reference (docs.kernel.org)
• Linux Kernel Documentation — UIO: Userspace I/O driver guide
• Linux Kernel Documentation — Linux Security Module (LSM) framework
• kernel.org — Official Linux kernel releases
• Oracle Blogs — Exploring Linux kernel configurations with kconfigs