How to set up a separate X server for the Oculus Rift?

Question

Oculus SDK version 0.4.3+ comes with an experimental Linux support for the Oculus Rift DK2. The documentation mentions two general options to run the Rift:

1. running the Rift in a twinview configuration using xrandr (or one of its graphical front-ends)
2. running the Rift as a separate X server.

My question is: What are the required steps to configure a separate X server for the Rift? In particular:

• Apparently, modern X11 versions no longer have a xorg.conf file. Is creating a xorg.conf file actually required? Or is it possible to make the necessary adjustments in /usr/share/X11/xorg.conf.d/ or /etc/X11/xorg.conf.d/, as suggested by this answer?
• If it is necessary to create an xorg.conf: How do I obtain and exactly replicate all non-Rift-related settings of my current default configuration?
• If I understand the xorg.conf man page correctly, it is required to define a new ServerLayout section for the Rift, since a ServerLayout is the top level definition of every X server. The ServerLayout itself binds a Screen and an InputDevice section. The Screen section further binds a Device section (a graphics device) and a Monitor section. Therefore, we probably need at least a SeverLayout, InputDevice, Device, and Monitor section. How would these sections typically look like for an Oculus Rift X server?