}

void TF_AttrBuilderCheckCanRunOnDevice(TF_AttrBuilder* builder,
                                       const char* device_type,
                                       TF_Status* status) {
  status->status = tensorflow::FindKernelDef(
      tensorflow::DeviceType(device_type), builder->BuildNodeDef(),
      /* def = */ nullptr, /* kernel_class_name = */ nullptr);
}

    if (n > INT32_MAX)
      requested_read_length = INT32_MAX;
    else
      requested_read_length = n;

    // `pread` returns a `ssize_t` on POSIX, but due to interface being
    // cross-platform, return type of `Read` is `int64_t`.
    int64_t r = int64_t{pread(posix_file->fd, dst, requested_read_length,
                              static_cast<off_t>(offset))};
    if (r > 0) {
      dst += r;

  std::string device_type = parsed_name.type;
  int device_id = 0;
  if (parsed_name.has_id) {
    device_id = parsed_name.id;
  }

  ctx.device_id = device_id;
  if (device_type == "CPU") {
    ctx.device_type = DLDeviceType::kDLCPU;
  } else if (device_type == "GPU") {
#if TENSORFLOW_USE_ROCM
    ctx.device_type = DLDeviceType::kDLROCM;
#else

  TF_RETURN_IF_ERROR(ValidateTPOptimizerConfigs(optimizer_configs));

  CGraphOptimizerRegister(
      [=]() { return new CGraphOptimizer(optimizer, params.device_type); },
      optimizer_configs, params.device_type);

  return absl::OkStatus();
}

}  // namespace grappler
}  // namespace tensorflow

void TF_GetNodesToPreserveListSize(const TF_GrapplerItem* item, int* num_values,

      gtl::ArraySlice<AbstractTensorHandle*> gradient_tensors) const override;

  // Calls the passed-in backward function.
  // op_type is the op's name provided in RecordOperation.
  Status CallBackwardFunction(
      const string& op_type, GradientFunction* gradient_function,
      const std::vector<int64_t>& unneeded_gradients,
      gtl::ArraySlice<AbstractTensorHandle*> output_gradients,

    return false;
  }
  bool found_dtype = false;
  bool found_shape = false;
  for (const auto& attr : node_def.attr()) {
    if (attr.first == "dtype") {
      if (attr.second.type() == tensorflow::DT_INT32) {
        found_dtype = true;
      } else {
        return false;
      }
    } else if (attr.first == "shape") {
      found_shape = true;
    }
  }

  executed = false;
  TFE_Execute(matmul.get(), &retval, &num_retvals, status.get());
  EXPECT_FALSE(executed);
  ASSERT_FALSE(TF_GetCode(status.get()) == TF_OK);

  // Custom devices can refuse to do type-based dispatch (as hcustom1 is
  // configured to do)
  matmul.reset(MatMulOp(context.get(), hcustom1.get(), hcpu.get()));
  num_retvals = 1;
  executed = false;

      tensorflow::DeviceType(device_type), std::move(*pjrt_client));
  status->status = s;
}

void TF_ResetPjRtCClient(const char* device_type, TF_Status* status) {
  status->status =
      tensorflow::ResetPjRtClient(tensorflow::DeviceType(device_type));
}

PJRT_Client* TF_GetPjRtCClient(const char* device_type, TF_Status* status) {
  absl::StatusOr<xla::PjRtCApiClient*> pjrt_c_api_client =

      "    signature {"
      "      name: 'AddVariablesFunction'"
      "      input_arg {"
      "        name: 'var'"
      "        type: DT_RESOURCE"
      "      }"
      "      output_arg {"
      "        name: 'sum'"
      "        type: DT_FLOAT"
      "      }"
      "    }"
      "    node_def {"
      "      name: 'read0'"
      "      op: 'ReadVariableOp'"
      "      input: 'var'"

    TF_SetStatus(status, TF_OK, "");
    PopulateDefaultParam(params);
    params->device_type = "Device0";
  };
  auto plugin_init_1 = [](TP_OptimizerRegistrationParams* const params,
                          TF_Status* const status) -> void {
    TF_SetStatus(status, TF_OK, "");
    PopulateDefaultParam(params);
    params->device_type = "Device1";
  };

  TF_ASSERT_OK(InitGraphPlugin(plugin_init_0));