The effort to merge the real-time preemption patch into the mainline kernel requires to refactor existing infrastructure in the kernel. Such refactoring work is intrusive and potentially disruptive.

This talk will take a close look at recent refactoring work and explain the scope, the benefits and the pitfalls. This includes a deep dive into some of the oldest parts of the kernel and their historical evolvement.

Thomas Gleixner, Linutronix