[llvm] GlobalISel: Relax verifier between physreg and typed vreg (PR #159281)

[Matt Arsenault via llvm-commits]

arsenm wrote:

> `%0:(s1) = COPY $vcc` looks more convenient then `s64`, why is copy better then doing `%0:(s1) = call i1 @llvm.amdgcn.inverse.ballot.i64 $vcc` Why do we need to use $vcc explicitly in GlobalISel?

Not vcc specifically, but this is to support lowering i1 argument values to lanemask SGPRs

https://github.com/llvm/llvm-project/pull/159281