[lld] [LLD] [COFF] [test] Readd lto-late-arm.ll (PR #143494)

[Matt Arsenault via llvm-commits]

arsenm wrote:

All of the cases set in ARMISelLowering (e.g. https://github.com/llvm/llvm-project/blob/9d491bc602c2d9730cb42fe25f0753471a3af389/llvm/lib/Target/ARM/ARMISelLowering.cpp#L683) are in danger of the same issue, should there be a new test that covers all of these cases? 

I'm working on making sure the full set is visible in RuntimeLibcalls and the lowering will then make a selection from that superset 

https://github.com/llvm/llvm-project/pull/143494