[SDAG] Allow scalable vectors in SimplifyDemanded routines

[reames]

Summary

This is a continuation of the series of patches adding lane wise support for scalable vectors in various knownbit-esq routines.

The basic idea here is that we track a single lane for scalable vectors which corresponds to an unknown number of lanes at runtime. This is enough for us to perform lane wise reasoning on many arithmetic operations.

[reames]

rebase and ping

[craig.topper]

llvm/lib/CodeGen/SelectionDAG/TargetLowering.cpp

2705

Have you already audited SimplifyDemandedBitsForTargetNode for scalable vectors?

[reames]

llvm/lib/CodeGen/SelectionDAG/TargetLowering.cpp

2705

Nope, good catch.

[reames]

address reviewer comment and add conservative bailout for target nodes

[reames]

ping

[craig.topper]

LGTM