ARM: Emit normal IR for vaddhn/vsubhn NEON intrinsics

ARM: Emit normal IR for vaddhn/vsubhn NEON intrinsics

These operations "vector add high-half narrow" actually correspond to the
sequence:

    %sum = add <4 x i32> %lhs, %rhs
    %high = lshr <4 x i32> %sum, <i32 16, i32 16, i32 16, i32 16>
    %res = trunc <4 x i32> %high to <4 x i16>

Now that LLVM can spot this, Clang should emit the corresponding LLVM IR.

git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@189463 91177308-0d34-0410-b5e6-96231b3b80d8