Linux kernel
Contributing
I have always been drawn to system programming and to understanding what happens further down the stack, close to the OS and hardware. Contributing to the Linux kernel grew out of that.
Getting started was the hardest part. The kernel is huge and carries decades of context, so at first it felt like a closed box. There wasn't one turning point where it suddenly made sense. I just started reading the code, chasing down bugs, and sending patches. Over time, once I found areas I was comfortable digging into, the codebase became much easier to navigate. Working on kernel-internals.org helped a lot too, since writing about how different subsystems fit together pushed me to read parts of the tree I wouldn't have touched otherwise.
Write-ups
- Loading an X.509 certificate with an empty subject
add_key() refused X.509 certificates with an empty Subject, even though RFC 5280 allows them. The cause was in the kernel's ASN.1 compiler. - Writes to queue/scheduler failing with EBUSY
Writes to queue/scheduler failed with -EBUSY due to a tag-set-wide lock. The fix takes it for read and uses a per-queue busy flag.
Understanding
There is already plenty of documentation covering what the kernel does today, and the source code is right there. What is harder to find is how it got that way: the trade-offs behind a design, the alternatives that were tried, and how subsystems evolved over time.
That is why I started kernel-internals.org, which also includes "War Stories" pages that walk through specific bugs and CVEs in detail.
Find me
You can browse my commits in the kernel git history and my mailing-list threads on lore.