[llvm] ISel: introduce vector ISD::LRINT, ISD::LLRINT; custom RISCV lowering (PR #66924)

[Craig Topper via llvm-commits]

================
@@ -2687,13 +2687,14 @@ static RISCVFPRndMode::RoundingMode matchRoundingOp(unsigned Opc) {
   return RISCVFPRndMode::Invalid;
 }
 
-// Expand vector FTRUNC, FCEIL, FFLOOR, FROUND, VP_FCEIL, VP_FFLOOR, VP_FROUND
-// VP_FROUNDEVEN, VP_FROUNDTOZERO, VP_FRINT and VP_FNEARBYINT by converting to
-// the integer domain and back. Taking care to avoid converting values that are
+// Expand vector FTRUNC, FCEIL, FFLOOR, FROUND, FRINT, VP_FCEIL, VP_FFLOOR,
+// VP_FROUND, VP_FROUNDEVEN, VP_FROUNDTOZERO, VP_FRINT and VP_FNEARBYINT by
+// converting to the integer domain and back. Expand LRINT and LLRINT by
+// converting to integer domain. Take care to avoid converting values that are
 // nan or already correct.
 static SDValue
-lowerVectorFTRUNC_FCEIL_FFLOOR_FROUND(SDValue Op, SelectionDAG &DAG,
-                                      const RISCVSubtarget &Subtarget) {
+lowerVectorFTRUNC_FCEIL_FFLOOR_FROUND_XRINT(SDValue Op, SelectionDAG &DAG,
----------------
topperc wrote:

Why are we sending LRINT and LLRINT through this code if we have to skip the majority of it? Can we just have a purpose built function for these?

https://github.com/llvm/llvm-project/pull/66924