From what I have gathered from different forums and bug lists is that this is not directly a driver issue but an issue between the driver and kernel. It seems to occur with different video drivers. On one forum, it was reported by someone that they compiled their own kernel without CFS support and all their freezups were gone.
Is this CFS Completely Fair Scheduler? I thought it would go into the kernel at 2.6.23? I hate to compile the kernel. Haven't done that for many years. Maybe I'll just stick with the nv driver, if my machine stops freezing. DS