[llvm] [Analysis] Attribute Range should not prevent tail call optimization (PR #91122)

[Matt Arsenault via llvm-commits]

arsenm wrote:

> declare range(i64 0, 2) i64 @llvm.llround.f32(float) nounwind readnone
> define i64 @testmsxs(float %x) {
>   entry:
>     %0 = musttail call i64 @llvm.llround.f32(float %x)
>       ret i64 %0
> }
> ```

I'm surprised we allow musttail on intrinsic calls, which aren't really supposed to have ABI 

https://github.com/llvm/llvm-project/pull/91122