[llvm] [SelectionDAG] Fold constant min/max vector reductions (PR #209190)

[Luke Lau via llvm-commits]

================
@@ -7802,18 +7808,25 @@ SDValue SelectionDAG::FoldConstantArithmetic(unsigned Opcode, const SDLoc &DL,
     if (Opcode == ISD::BITCAST)
       return SDValue();
 
-    // Constant fold VECREDUCE_ADD with a BUILD_VECTOR of integer constants.
-    if (Opcode == ISD::VECREDUCE_ADD && ISD::isBuildVectorOfConstantSDNodes(N1.getNode())) {
+    // Constant fold integer vector reductions with constant BUILD_VECTORs.
+    if ((Opcode == ISD::VECREDUCE_ADD || Opcode == ISD::VECREDUCE_SMAX ||
+         Opcode == ISD::VECREDUCE_SMIN || Opcode == ISD::VECREDUCE_UMAX ||
+         Opcode == ISD::VECREDUCE_UMIN) &&
+        ISD::isBuildVectorOfConstantSDNodes(N1.getNode())) {
       unsigned EltBits = N1.getValueType().getScalarSizeInBits();
-      APInt Acc = APInt::getZero(EltBits);
+      EVT EltVT = N1.getValueType().getScalarType();
----------------
lukel97 wrote:

Nit I think this line in the diff isn't needed anymore, it can be moved back to line 7830

https://github.com/llvm/llvm-project/pull/209190