Question about Value Range Propagation

Hello ,

I have came across this discussion on VRP in LLVM.
Is this implementation is in mainline LLVM ?
Is Patterson’s approach available or still work is required ?
It is Inter-procedural VRP or not ?

Sincerely,