But I wouldn't go for Common Lisp. I'd go for something smaller and cleaner.
1. The ability to manually manage memory instead of relying on a GC.
2. Low-level access to memory. As in, the ability to address an arbitrary location in physical memory
3. Inline assembly and the ability to link with assembly code
There have been operating systems written in Lisp in the past, such as the operating systems for the various Lisp Machines. https://en.wikipedia.org/wiki/Lisp_machine
In other words, the problem is not with Lisp, nor with a lack of support for low-level features, but with tooling. The tool support for writing a kernel in Common Lisp is pretty bad at the moment, and there are not many people trying to improve on that.
It's difficult to take the question seriously rather than as an attention-grab.
I think CL is in a way more operating system like than most other Lisps and would make good base for operating system. You can load and unload stuff, clean the system, etc.
Linux took many years to get to the point where it had the stability and performance to be able to compete with commercial operating systems. Users of Linux will not care what language the kernel is written in. They're only going to care whether it will run all of their existing applications without changes, as fast and reliably as Linux does. If the new system doesn't, it will never have any users. (People were willing to put up with the early versions of Linux because it was free and its competitors were not. Any new OS would have to compete with an existing free OS.)
Who would write and maintain it? Linus and most of his followers would stick with the C version, which they know and love, so where are we going to suddenly find the hundreds or thousands of contributors who are both expert Lisp programmers and expert kernel hackers?
I would add: if someone really _did_ do all the work to write a whole modern operating system in Lisp, it would be criminally unimaginative to use the Linux ABI as your interface to the world.
You would have to write your own userspace too for the whole enterprise to be worth bothering with. And then you have basically written SqueakNOS (http://squeaknos.blogspot.ch). :-)
For many of the same reasons people started writing OS kernels in C rather than assembly: Less scut work, wrong code looks wrong more often, and more opportunity for code reuse.
Other reasons include being able to provide a higher-level API, which only kinda happened with the move to C. A Lisp API would enable things like system-level garbage collection and possibly things like a single-address-space OS where the kernel is responsible for fine-grained memory protection and multiprocessing becomes as cheap as multithreading without becoming as unsafe as multithreading.
But that wouldn't be Linux anymore, I suppose; it would be a Linux-compatible system with extended features beyond what the C kernel provides. On the other hand, is Linux 3.11 the same as Linux 1.0 or is it merely mostly-compatible with Linux 1.0?
That's a one man effort to write a Lisp OS for x86 from scratch. He wrote his own compiler, basically with the chain of macroexpansions bottoming out in WITH-INLINE-ASSEMBLER.
I only played with it once. I was struck by how cool it was that I could play with the Lisp REPL from inside the garbage collector (also written in Lisp). This was surprising but then seemed obvious: the GC is simply a Lisp function that's moving objects from old-space into new-space (and then deleting old-space), and it doesn't mind if you wake up and interleave a few more allocations into new-space.
Movitz was current back in 2004 when Pascal Costanza organized the first real european Lisp workshop in Oslo and a whole bunch of Lisp hackers who only knew each other through the 'net got to meet up in person. fun times.
