[Mlir-commits] [mlir] [mlir][ArithToSPIRV] Fix crash on tensor arith.uitofp/sitofp lowering (PR #227102)

[Igor Wodiany]

================
@@ -731,6 +731,14 @@ struct IntToFPPattern final : public OpConversionPattern<ArithOp> {
     // Check if the source integer type was widened during type conversion.
     unsigned originalBitwidth =
         getElementTypeOrSelf(op.getIn().getType()).getIntOrFloatBitWidth();
+
+    if (!srcType ||
+        !(isa<IntegerType, FloatType>(getElementTypeOrSelf(srcType)) ||
+          isa<VectorType>(srcType))) {
----------------
IgWod wrote:

Do we need this part of the check? If we hit it, it would mean element type is neither int nor float and that would still allow the code below to crash, no?

https://github.com/llvm/llvm-project/pull/227102