This repository contains all the needed pieces to use OpenGL on a headless
Nvidia GPU (such as in Optimus configuration) on FreeBSD, without interfering
with functionality of primary GPU device (such as Intel).

Optionally, graphical output of application running on Nvidia can be forwarded
to primary display using VirtualGL.

The ports are designed for eventual integration into FreeBSD ports tree.



Testing / installation:

Assumes $CHECKOUT is where this project is checked out
Assumes https://svn.freebsd.org/ports/head is checked out to /usr/ports

Setup - as root:

  Prepare ports tree

        svn patch $CHECKOUT/ports/x11_nvidia-driver.diff /usr/ports
        svn patch $CHECKOUT/ports/x11_nvidia-settings.diff /usr/ports
        cp -r $CHECKOUT/ports/x11_nvidia-headless-driver \
          /usr/ports/x11/nvidia-headless-driver
        cp -r $CHECKOUT/ports/x11_nvidia-headless-utils \
          /usr/ports/x11/nvidia-headless-utils
        cp -r $CHECKOUT/ports/x11_nvidia-hybrid-graphics \
          /usr/ports/x11/nvidia-hybrid-graphics

  Build ports

        make -C /usr/ports/x11/nvidia-hybrid-graphics install

Startup - as root:

        sysrc nvidia_xorg_enable=YES
        service nvidia_xorg start

Running applications - as user:

  (headless)
        nvrun glxinfo -B

  (graphical output):
        nvrun-vgl glxgears