Use OpConversionPattern instead of BufferizeOpConversionPattern

As described in mlir/Transforms/Bufferize.h, patterns that don't need the special methods on a BufferizeTypeConverter should use a regular OpConversionPattern.

PiperOrigin-RevId: 338424819

lib/Dialect/mhlo/transforms/hlo_legalize_to_lhlo.cc

@@ -42,7 +42,7 @@ namespace mhlo {
 namespace {
 
 template <typename T>
-using BaseOpConversion = BufferizeOpConversionPattern<T>;
+using BaseOpConversion = OpConversionPattern<T>;
 
 Value InsertDynamicAllocAndDealloc(Location loc, Value result,
                                    Value shape_operand,
@@ -623,7 +623,7 @@ void populateHLOToLHLOConversionPattern(MLIRContext* context,
       HloToLhloReturnOpConverter,
       HloToLhloTensorLoadOpConverter,
       HloToLhloTensorStoreOpConverter
-  >(context, *converter);
+  >(context);
   // clang-format on
 }