It is not really about reserving cores.

Just that - even if you distribute work to a number of cores (that itself limited), you still need to sync the execution. So unless you have some bigger chunks of computation, the performance increase will be very small.

And yes, there are bigger chunks of work when rendering a web page, but that is typically already outside the css engine and can be distributed without having to rewrite the core engine.

But my comments are a bit generic here, so I am open to comments regarding this particular project.

Servo also had some other novel architectural ideas IIRC, not directly related to paralelism, but it's been a while since I was looking into that, I don't remember much.