[clang] [clang][SPIR-V][AMDGPU] Fix variadic aggregate ABI classification for AMDGCN (PR #216326)

[Matt Arsenault via cfe-commits]

arsenm wrote:

> > It would be good if we could move this code into the ABI library
> 
> Thanks for noting! Moved, LoC ratio became much better

I'd rather do the move and bug fix separately, in either order 

(I eventually need the ABI library support to be complete for libcall emission improvements)

https://github.com/llvm/llvm-project/pull/216326