Why can you set the stack size on CUDA GPU devices with compute capability 2.0+ only?

Why can you set the stack size on CUDA GPU devices with compute capability
2.0+ only?

I guess I have a couple questions related to the call stack, but my main
one would be,
"Why can you set the stack size on devices with compute capability 2.0+
only?"
Also, why can I only view the call stack using nvidia Nsight for devices
with compute capability 2.0 and above only? What's so special about 2.0+
devices that allows them to do this, but not 1.x devices?
I know this is a pretty broad question, but any insight into the above
questions' answers or even information on the stack call on CUDA devices
in general would be highly appreciated.