As Linux - Linux/x86 does not make use of "corner cases" which require complex software techniques (e.g. dynamic translat I/O and other execut I/O monitoring methods), plex86 now implements no such techniques.   Guest Linux code is executed in a safe VM environment managed by a Virtual Machine - VM monitor.   Guest kernel code is executed at a less priviliged level, causing privileged operat I/Ons to generate except I/Ons which are managed by the VM monitor.

Virtualized I/O model:  because the Linux kernel can easily be configured before compiling to remove support for most all I/O hardware, the model for handling I/O in the guest Linux VM is greatly simplified and improved in plex86.   Excepting for some very lightweight emulatI/On of a few core components (e.g. interupt controller, system timer), plex86 does not model(emulate) I/O hardware.   All I/O interactI/Ons occur through a Hardware AbstractI/On Layer (HAL) which passes packets of informatI/On between the guest Linux and the host OS.   This not only removes the complexity/overhead of modeling I/O hardware, but allows for more flexibility in dealing with guest-host interactI/Ons and creates a truly virtualized Linux resource which is unrooted from the actual hardware.   Guest Linux drivers interact with the HAL to manage I/O related interactI/Ons (networking/disk/etc) with the host OS.