break;
      case FetchState::FINISHED:
        return TF_SetStatus(status, TF_OK, "");
    }
  }
  return TF_SetStatus(
      status, TF_INTERNAL,
      "Control flow should never reach the end of RamFileBlockCache::Fetch.");
}

int64_t RamFileBlockCache::Read(const std::string& filename, size_t offset,
                                size_t n, char* buffer, TF_Status* status) {
  if (n == 0) {

  // The cache has space for `block_count` blocks. The loop with i = 0 should
  // fill the cache, and the loop with i = 1 should be all cache hits. The
  // fetcher checks that it is called once and only once for each offset (to
  // fetch the corresponding block).
  for (int i = 0; i < 2; i++) {
    for (int j = 0; j < block_count; j++) {
      TF_EXPECT_OK(ReadCache(&cache, "", block_size * j, block_size, &out));
    }
  }

    TF_Tensor* f_t = TFE_TensorHandleResolve(handle, s);
    ASSERT_EQ(TF_OK, TF_GetCode(s)) << TF_Message(s);

    memcpy(&result_data[0], TF_TensorData(f_t), TF_TensorByteSize(f_t));

    // Verify results for each output
    for (int j = 0; j < 4; j++) {
      ASSERT_EQ(result_data[j], expected_outputs[idx][j]);
    }

    TF_DeleteTensor(f_t);
  }

  // Free memory associated with add and MatMul outputs

    change.
-   You make the change and submit it for the review again.
-   This cycle repeats itself until the PR gets approved.
-   Note: As a friendly reminder, we may reach out to you if the PR is awaiting
    your response for more than 2 weeks.

**4. Approved**

-   Once the PR is approved, it gets `kokoro:force-run` label applied and it
    initiates CI/CD tests.

      on behalf of whom a Contribution has been received by Licensor and
      subsequently incorporated within the Work.

   2. Grant of Copyright License. Subject to the terms and conditions of
      this License, each Contributor hereby grants to You a perpetual,
      worldwide, non-exclusive, no-charge, royalty-free, irrevocable
      copyright license to reproduce, prepare Derivative Works of,

        "memory regions");

  return OkStatus();
}

// Validates the operations supplied by the plugin.
//
// Uses the 4 simpler `ValidateHelper(const TF_...*)` to validate each
// individual function table and then checks that the function table for a
// specific file type exists if the plugin offers support for creating that
// type of files.

    // Get transpose attrs
    bool t_a;
    TF_RETURN_IF_ERROR(forward_attrs_.Get("transpose_a", &t_a));

    bool t_b;
    TF_RETURN_IF_ERROR(forward_attrs_.Get("transpose_b", &t_b));

    // Conj each input
    AbstractTensorHandle* conj_output;
    std::string name = "Conj_A_MatMul_Grad";
    TF_RETURN_IF_ERROR(
        SafeConj(ctx, forward_inputs_[0], &conj_output, name.c_str()));

  // (as in the case of ConstOp kernel creation on GPU, which involves copying a
  // CPU tensor to GPU).
  // Setting a larger thread pool does not help with the Swift caller, as we use
  // a different TFE context for each thread of execution (for running graph
  // functions, and their send/recvs corountines).
  config.set_inter_op_parallelism_threads(1);

  TF_Buffer* ret = TF_NewBuffer();
  TF_CHECK_OK(MessageToBuffer(config, ret));

//
// First we filter the tape to just the subset of operations we want to
// differentiate. In the process of doing so we count how many times each Tensor
// is used as an input to an op (so we know when we're done computing gradients
// for that Tensor). We also count, for each tape entry, how many of its output
// Tensors need gradients to be computed (Tensors which are not used do not need
// any gradients to be computed).
//

  // A non-blocking version of `Execute`. After each call, `Join` must be called
  // before `StartExecute` is called again. Using `StartExecute` with `Join`
  // allows the caller to schedule computation on multiple ParallelDevices
  // without sequencing those operations (first call `StartExecute` on each
  // parallel device, then call `Join` on each; even if some of the `Join`s