This now makes the scripts more compact as there are conditional statements
selecting the vGPU profile enabled card by the generation, and the generation
is determined by the ranges of PCI IDs. This should eliminate the need for
adding PCI IDs manually now, and open up more profile options for users to pick
from, including profiles with a VRAM multiple of 3GB.
Note that this breaks compatibility with previous versions as users may have
to go back and reselect a new profile.
The previously unknown members of vgpu_unlock_vgpu_t are now known. They decode
to an additional name for the vGPU and the length of the vGPU names. The code
has been updated to provide correct values for these members.
Some ioctl calls result in "Assertion Failed" although everything seems to work
anyway. So we suppress these so that the error messages that do pop up are
actually significant.
Also add support for sending syslog messages from the vgpu_unlock python script.
This simplifies debugging a lot, and allows us to log the operation type values
for each ioctl call that fails.
