[llvm] SelectionDAG: Support nofpclass(nan/qnan/snan/nzero) in arguments (PR #130051)

[Matt Arsenault via llvm-commits]

================
@@ -11827,10 +11827,25 @@ void SelectionDAGISel::LowerArguments(const Function &F) {
           AssertOp = ISD::AssertSext;
         else if (Arg.hasAttribute(Attribute::ZExt))
           AssertOp = ISD::AssertZext;
-
-        ArgValues.push_back(getCopyFromParts(DAG, dl, &InVals[i], NumParts,
-                                             PartVT, VT, nullptr, NewRoot,
-                                             F.getCallingConv(), AssertOp));
+        SDValue OutVal =
+            getCopyFromParts(DAG, dl, &InVals[i], NumParts, PartVT, VT, nullptr,
+                             NewRoot, F.getCallingConv(), AssertOp);
+        if (Arg.hasAttribute(Attribute::NoFPClass) &&
+            OutVal.getValueType().isFloatingPoint()) {
+          SDNodeFlags OutValFlags = OutVal->getFlags();
+          bool NoSNaN = ((Arg.getNoFPClass() & llvm::fcSNan) == llvm::fcSNan);
----------------
arsenm wrote:

AssertNoFPClass should have an explicit integer mask argument that directly copies the value. You shouldn't be trying to re-encode that value as fast math flags 

https://github.com/llvm/llvm-project/pull/130051