#include "tensorflow/c/experimental/stream_executor/stream_executor_internal.h"
#include "tensorflow/c/tf_status_helper.h"
#include "xla/stream_executor/executor_cache.h"
#include "xla/stream_executor/host_memory_allocation.h"
#include "xla/stream_executor/memory_allocation.h"
#include "xla/stream_executor/platform.h"
#include "xla/stream_executor/platform_manager.h"
#include "xla/stream_executor/stream.h"

#include "xla/stream_executor/executor_cache.h"
#include "xla/stream_executor/platform.h"
#include "xla/stream_executor/stream_common.h"
#include "xla/stream_executor/stream_executor.h"
#include "tsl/platform/statusor.h"

namespace stream_executor {

// Plugin initialization function that a device plugin
// must define.
typedef void (*SEInitPluginFn)(SE_PlatformRegistrationParams* const,
                               TF_Status* const);

#include "tsl/platform/statusor.h"

namespace tensorflow {

// The allocator used for Tensors assigned to the XLA device.
XlaDeviceAllocator::XlaDeviceAllocator(
    stream_executor::StreamExecutor* stream_executor)
    : stream_executor_(stream_executor) {}

XlaDeviceAllocator::~XlaDeviceAllocator() = default;

std::string XlaDeviceAllocator::Name() { return "xla"; }

#include "tensorflow/c/experimental/stream_executor/stream_executor.h"

#include <functional>
#include <utility>

#include "tensorflow/c/experimental/stream_executor/stream_executor_internal.h"
#include "tensorflow/c/experimental/stream_executor/stream_executor_test_util.h"
#include "xla/stream_executor/event.h"
#include "xla/stream_executor/platform_manager.h"
#include "xla/stream_executor/stream.h"

#include "tensorflow/compiler/tf2xla/shape_util.h"
#include "tensorflow/compiler/tf2xla/xla_op_registry.h"
#include "xla/stream_executor/device_memory.h"
#include "xla/stream_executor/platform_manager.h"
#include "xla/stream_executor/stream.h"
#include "xla/stream_executor/stream_executor.h"
#include "tensorflow/core/framework/tensor_testutil.h"
#include "tsl/lib/core/status_test_util.h"

namespace tensorflow {
namespace {

)

cc_library(
    name = "stream_executor_hdrs",
    hdrs = ["stream_executor.h"],
    visibility = ["//tensorflow:internal"],
    deps = [
        "//tensorflow/c:c_api_macros_hdrs",
        "//tensorflow/c:tf_status_headers",
    ],
)

cc_library(
    name = "stream_executor",
    srcs = ["stream_executor.cc"],
    hdrs = ["stream_executor.h"],
    visibility = ["//tensorflow:internal"],
    deps = [

#include "tensorflow/compiler/tf2xla/xla_compiler.h"
#include "xla/client/client_library.h"
#include "xla/client/compile_only_client.h"
#include "xla/stream_executor/host/host_platform_id.h"
#include "xla/stream_executor/platform_manager.h"
#include "tensorflow/core/common_runtime/graph_constructor.h"
#include "tensorflow/core/framework/function.h"
#include "tensorflow/core/framework/graph.pb.h"

        "@local_tsl//tsl/platform:statusor",
        "@local_xla//xla/stream_executor/tpu:c_api_conversions",
        "@local_xla//xla/stream_executor/tpu:status_helper",
        "@local_xla//xla/stream_executor/tpu:tpu_api",
        "@local_xla//xla/stream_executor/tpu:tpu_executor_base",
        "@local_xla//xla/stream_executor/tpu:tpu_node_context",
        "@local_xla//xla/stream_executor/tpu:tpu_platform_interface",

#include "tensorflow/compiler/mlir/tensorflow/ir/tf_ops.h"
#include "xla/layout.h"
#include "xla/mlir_hlo/mhlo/IR/hlo_ops.h"
#include "xla/shape.h"
#include "xla/stream_executor/tpu/c_api_conversions.h"
#include "xla/stream_executor/tpu/tpu_api.h"
#include "xla/translate/mhlo_to_hlo/type_to_shape.h"

namespace mlir {

static FailureOr<std::vector<int64_t>> GetTPUInfeedLayoutFromAPI(
    RankedTensorType t) {

#ifndef TENSORFLOW_COMPILER_JIT_XLA_HOST_SEND_DEVICE_CONTEXT_H_
#define TENSORFLOW_COMPILER_JIT_XLA_HOST_SEND_DEVICE_CONTEXT_H_

#include "xla/shape.h"
#include "xla/stream_executor/device_memory.h"
#include "xla/stream_executor/stream.h"
#include "tensorflow/core/framework/device_base.h"
#include "tfrt/concurrency/async_value_ref.h"  // from @tf_runtime

namespace tensorflow {