[LLVMdev] [AArch64] Should we restrict to the pointer type used in ldN/stN intrinsics?

[Tim Northover]

Hi Hao,

> I just think it’s strange that the pointer has no relationship with the
> returned type. Currently there are IR regression test cases using different
> kinds of pointers like ‘xx.ld2.v4i32.p0i32’, ‘xx.ld2.v4i32.p0v4i32’ or
> ‘xx.ld2.v4i32.p0i8’, which looks confusing. Should we modify the definition
> of such intrinsics and restrict the pointer type?

I'm not sure I'd bother. We're hoping to move to a single pointer type
throughout LLVM, and this sounds like something that would just have
to be undone when we do. On the other hand, there's not been much
actual progress there for a while, at least that I've noticed.

Is is causing actual problems that you've seen, or is it a cleanliness goal?

Cheers.

Tim.