[llvm] [SystemZ] Allow forming overflow op for i128 (PR #153557)

[Ulrich Weigand via llvm-commits]

================
@@ -523,7 +523,7 @@ class SystemZTargetLowering : public TargetLowering {
                             bool MathUsed) const override {
     // Form add and sub with overflow intrinsics regardless of any extra
     // users of the math result.
-    return VT == MVT::i32 || VT == MVT::i64;
+    return VT == MVT::i32 || VT == MVT::i64 || VT == MVT::i128;
----------------
uweigand wrote:

That looks fine for z13 and higher, but on older machines, `i128` is not actually a legal type, and we have no instructions for 128-bit operflowing ops.   The common code `shouldFormOverflowOp` has a `!isOperationExpand(Opcode, VT)` check - I think we need that here for the `i128` case.

https://github.com/llvm/llvm-project/pull/153557