const string& got_filename, size_t got_offset,
                     size_t got_n, char* buffer, TF_Status* status) -> int64_t {
    EXPECT_EQ(got_filename, want_filename);
    EXPECT_EQ(got_offset, want_offset);
    EXPECT_EQ(got_n, want_n);
    calls++;
    memset(buffer, 'x', got_n);
    TF_SetStatus(status, TF_OK, "");
    return got_n;
  };

      dynamic_addr(symbol) = static_addr(symbol) + load_addr
 
-     The choice of symbol is arbitrary. The static address we obtain
-     by constructing a non GOT reference to the symbol, the dynamic
-     address of the symbol we compute using adrp/add to compute the
-     symbol's address relative to the PC. */
-
-  ElfW(Addr) static_addr;
-  ElfW(Addr) dynamic_addr;
-

   *     my_add2 = b + b
   *     return my_add1, my_add2
   *
   * Now we will execute this function with an eager context:
   *
   *   output1, output2 = two_adds(2.0, 3.0)
   *
   * and check that we got 5.0 and 6.0 as results.
   */

  // Build eager context.
  TFE_ContextOptions* opts = TFE_NewContextOptions();
  TFE_ContextOptionsSetTfrt(opts, std::get<1>(GetParam()));
  TF_ExecutionContext* eager_execution_ctx =

  switch (dtype.code) {
    case DLDataTypeCode::kDLBool:
      if (dtype.bits != 8) {
        return tensorflow::errors::InvalidArgument(
            "Only DLPack bools of bitwidth 8 are supported, got: ", dtype.bits);
      }
      *tf_dtype = TF_DataType::TF_BOOL;
      return absl::OkStatus();

    case DLDataTypeCode::kDLUInt:
      switch (dtype.bits) {
        case 8:

      std::string message(absl::StrCat(
          "The parallel device ", parallel_device_name, " expected ",
          parallel_device.num_underlying_devices(),
          " inputs to TPUReplicatedInput, but got ", inputs.size()));
      TF_SetStatus(status, TF_INVALID_ARGUMENT, message.c_str());
      return result;
    }
    std::vector<TensorHandlePtr> components;
    components.reserve(inputs.size());

    ASSERT_EQ(nodes.size(), fdef.node_def_size())
        << "Got unexpected number of nodes. Expected: ["
        << absl::StrJoin(nodes, ", ")
        << "] Actual nodes in fdef: " << fdef.DebugString();
    for (const NodeDef& node_def : fdef.node_def()) {
      ASSERT_TRUE(nodes.find(node_def.name()) != nodes.end())
          << "Got unexpected node: " << node_def.name()
          << " in fdef: " << fdef.DebugString();

  } else if (!DataTypeIsComplex(BaseType(dtype)) &&
             BaseType(dtype) != DT_VARIANT) {
    return errors::InvalidArgument(
        "Expected numeric or variant tensor, got dtype ", dtype);
  }
  return tensorflow::ops::Conj(ctx, input, output, name);
}

class AddGradientFunction : public GradientFunction {
 public:
  Status Compute(AbstractContext* ctx,

        tensorflow::down_cast<const tensorflow::TensorInterface*>(src->tensor);

    if (dst->dims() != 0) {
      return InvalidArgument(
          "Malformed TF_RESOURCE tensor: expected a scalar, got a tensor with "
          "shape ",
          dst->shape().DebugString());
    }
    *dst = tensorflow::Tensor(tensorflow::DT_RESOURCE, dst->shape());
    if (!dst->scalar<tensorflow::ResourceHandle>()().ParseFromString(

    status->status =
        tensorflow::errors::InvalidArgument("Unparseable AttrValue proto");
    return;
  }
  if (op == nullptr) {
    status->status = tensorflow::errors::InvalidArgument(
        "Got a null or uninitialized `op` argument");
    return;
  }
  tensorflow::EagerOperation* operation =
      OperationFromInterface(tensorflow::unwrap(const_cast<TFE_Op*>(op)));

    if (read >= 0) content.resize(read);
    if (file_size != content.size())
      TF_SetStatus(
          status_, TF_DATA_LOSS,
          std::string("expected " + std::to_string(file_size) + " got " +
                      std::to_string(content.size()) + " bytes")
              .c_str());
    return content;
  }

 protected:
  TF_Filesystem* filesystem_;
  TF_Status* status_;

 private: