Improvements:

- Emulate the instructions by executing them directly on the host CPU.
  This is easier and probably faster than doing it in software
  manually.

- Decode SUB from Primary, CMP from Group1, TEST from Group3, and add
  associated tests.

- Handle correctly the cases where an instruction that always implicitly
  reads the register operand is executed with the mem operand as source
  (eg: "orq (%rbx),%rax").

- Fix the MMU handling of 32bit-PAE. Under PAE CR3 is not page-aligned,
  so there are extra bits that are valid.

With these changes in place I can boot Windows XP on Qemu+NVMM.

(maxv)