[MKLDNN] NDArray reorder in C API and deconv (#16265)

* layout conversion in MXNDArrayGetData

* deconv: layout conversion for grad

* fix lint complain

src/c_api/c_api.cc

@@ -627,6 +627,13 @@ int MXNDArrayGetData(NDArrayHandle handle,
                      void **out_pdata) {
   API_BEGIN();
   NDArray *arr = static_cast<NDArray*>(handle);
+#if MXNET_USE_MKLDNN == 1
+  NDArray temp = *arr;
+  if (arr->IsMKLDNNData()) {
+    temp = arr->Reorder2Default();
+    arr = &temp;
+  }
+#endif
   if (!arr->is_none()) {
     *out_pdata = arr->data().dptr_;
   } else {

src/operator/nn/mkldnn/mkldnn_deconvolution.cc

@@ -559,10 +559,13 @@ void MKLDNNDeconvolutionBackward(const nnvm::NodeAttrs &attrs,
     Stream<cpu> *s = ctx.get_stream<cpu>();
     Tensor<cpu, 1, DType> gbias =
         in_grad[deconv::kBias].data().get<cpu, 1, DType>(s);
-    // If there is bias, the out grad has already been converted to the default
-    // format, so this shouldn't cause any performance issues.
-    Tensor<cpu, 4, DType> grad =
-        inputs[deconv::kOut].data().get<cpu, 4, DType>(s);
+
+    NDArray temp = inputs[deconv::kOut];
+    if (temp.IsMKLDNNData()) {
+      temp = temp.Reorder2Default();
+    }
+
+    Tensor<cpu, 4, DType> grad = temp.data().get<cpu, 4, DType>(s);
     Assign(gbias, req[deconv::kBias],
            mshadow::expr::sumall_except_dim<1>(grad));
   }