[clang] clang/CodeGen/RISCV: test lowering of math builtins (PR #71399)

[Ramkumar Ramachandra via cfe-commits]

artagnon wrote:

The main takeaways here:
- There is only ever an i64 variant of llrint produced, and we have tested codegen of that on RV32 and RV64.
- Only an i32 variant of lrint is produced on RV32, and an i64 variant on RV64. We have tested the codegen of both.

Hence, I would conclude that #70926 is closed justifiably, and #69945 is safe to re-land with respect to RISC-V (there have been some other fixes to it that have already landed).

https://github.com/llvm/llvm-project/pull/71399