compiler_rt: Update Windows ABI for float<->int conversion routines
Starting with LLVM 14, the Libcalls to these functions are now lowered
using a Vec(2, u64) instead of the standard ABI for i128 integers, so
our compiler-rt implementation needs to be updated to expose the same
ABI on Windows.
680419c407Cody Tapscott committed on 7/11/2022, 3:51:34 AM· committed by Andrew Kelleyparentf3333a5