Used to be my personal blog. Now, it covers programming, low-level development, operating systems, and more.

Nix Overlays: Add attribute to “lib” and avoid “infinite recursion error”

Today, I was about to add an attribute to the lib attribute of nixpkgs using a nixpkgs overlay. I thought that this is as straight forward as adding regular packages directly in an overlay. This wasn’t my first time using overlays, but my first time adding something to lib. There was a tiny but ugly pitfalls, that made me create this short blog post. Let’s look at the following code foo.nix: With I...

Read more

Writing a Linux Driver for QEMU’s Debugcon Device

In earlier blog posts, I discussed QEMU’s debugcon device and presented a minimal driver for it written in x86 assembly. In this blog post, we go a little higher and write a driver for Linux in C. The driver should provide a /dev/debugcon device so that processes can open the driver device file and write data to it. Hence, $ echo "hello from Linux" > /dev/debugcon will write to the debugcon device. Only...

Read more

Building an Out-of-Tree Linux Kernel Module in Nix

In this blog post, I’m going to show you briefly how you can compile an out-of-tree Linux kernel module in Nix. I use regular Nix for that and no Nix flakes. However, with flakes, the approach is similar. Background Knowledge An out-of-tree Linux kernel module is a standalone piece of code outside the Linux kernel source repository. They are used when people don’t want to, are not allowed to or can’t upstream drivers....

Read more

Nix: Testing a Single NixOS Module in CI

In this blog post, I explain how to test a single NixOS module in CI. By single NixOS module, I mean a module that provides some options and configurations, but itself is not a valid NixOS configuration. The use case is to have a common module (in a git repository) that you want to integrate in a full NixOS configuration, when you don’t have or want the full NixOS configuration at the same Read more…

By Philipp Schuster, ago

Kernel/Boot: Cope with a Relocation by a Bootloader in 32-bit x86 Assembly Code

A kernel should be relocatable by bootloaders as this is the only guarantee that it doesn’t interfere with addresses already in use by the firmware. This is especially true for the legacy boot flow on x86, i.e., without UEFI. TL;DR Introduction There are multiple ways to make a kernel relocatable. One option is to use the “relocatable” tag when the kernel uses the Multiboot2 boot flow. The assembly code must cope with runtime Read more…

By Philipp Schuster, ago

How Does the “File Size is Smaller Than Mem Size” Optimization Work in GNU ld for the .bss Section?

This article explains how the GNU linker (GNU ld) can save disk space for symbols in the .bss section. TL;DR: If a section is of type SHT_NOBITS and if it is the last section in a LOAD segment, GNU ld uses this optimization. When you build software with a compiled language, such as C, certain symbols, which may originate in variable declarations, may be initialized to zero. An example are the following two Read more…

By Philipp Schuster, ago