// RUN: tf-opt %s -split-input-file -verify-diagnostics -tf-xla-validate-inputs

// expected-error @+1 {{expects no nested calls of entry functions as they prevent graph traversal in some passes from working correctly}}
func.func @nested_entry_functions() attributes {tf.entry_function = {}} {
  tf_executor.graph {

        "@llvm-project//llvm:Support",
        "@llvm-project//mlir:FuncDialect",
        "@llvm-project//mlir:IR",
        "@llvm-project//mlir:Support",
        "@local_xla//xla/service:hlo_parser",
        "@local_xla//xla/service:hlo_proto_cc",
        "@local_xla//xla/translate/hlo_to_mhlo:hlo_to_mlir_hlo",
    ],
)

# Smaller version of flatbuffer_translate which only converts flatbuffer to MLIR.
cc_library(

#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/stream_executor.h"
#include "xla/stream_executor/stream_executor_common.h"

        [],
        otherwise = [
            "@local_xla//xla/stream_executor/cuda:all_runtime",
            "@local_xla//xla/stream_executor/cuda:cuda_driver",
            "@local_xla//xla/stream_executor/cuda:cuda_platform",
            "@local_xla//xla/stream_executor/cuda:cudnn_plugin",
            "@local_xla//xla/stream_executor/cuda:cufft_plugin",
            "@local_xla//xla/stream_executor:cuda_platform",
        ],
    ),

        "@llvm-project//mlir:QuantOps",
        "@llvm-project//mlir:Support",
        "@llvm-project//mlir:TranslateLib",
        "@local_xla//xla/hlo/ir:hlo",
        "@local_xla//xla/service:hlo_module_config",
        "@local_xla//xla/service:hlo_proto_cc",
        "@local_xla//xla/translate/mhlo_to_hlo:type_to_shape",
        "@stablehlo//:stablehlo_ops",
    ],
    alwayslink = 1,
)

cc_library(

      llvm::cl::desc("Choose target opset."),
      llvm::cl::values(
          clEnumValN(OpSet::TF, "TF",
                     "Uses TF ops that mimic quantization behavior"),
          clEnumValN(OpSet::XLA, "XLA", "Uses TF XLA ops"),
          clEnumValN(OpSet::UNIFORM_QUANTIZED, "UNIFORM_QUANTIZED",
                     "Uses TF Uniform Quantized ops"))};

  Option<int64_t> min_num_elements_for_weights_{

      llvm::cl::desc("Choose target opset."),
      llvm::cl::values(
          clEnumValN(OpSet::TF, "TF",
                     "Uses TF ops that mimic quantization behavior"),
          clEnumValN(OpSet::XLA, "XLA", "Uses TF XLA ops"),
          clEnumValN(OpSet::UNIFORM_QUANTIZED, "UNIFORM_QUANTIZED",
                     "Uses TF Uniform Quantized ops"))};
};

// Check if given indices in `val1` has same number of elements as given

CreateTPUClusterFormationPass(bool strict_clusters = false);

// Creates a pass that extracts outside compilation (Host ops inside device
// cluster) at head/tail of Device cluster to run before/after XLA computation.
std::unique_ptr<mlir::OperationPass<mlir::ModuleOp>>
CreateExtractHeadTailOutsideCompilationPass();

// Creates a pass that extract outside compilation (Host ops inside cevice

        "@local_tsl//tsl/platform:error_logging",
        "@local_tsl//tsl/platform:errors",
        "@local_xla//xla:xla_data_proto_cc",
        "@local_xla//xla:xla_proto_cc",
        "@local_xla//xla/mlir_hlo",
    ],
)

cc_library(
    name = "tpu_variable_runtime_reformatting",
    srcs = ["tpu_variable_runtime_reformatting.cc"],
    hdrs = [
        "runtime_passes.h",
    ],
    deps = [

#     --config=dbg --per_file_copt=+tensorflow/core/kernels/identity_op.*@-g
# Since this .bazelrc file is synced between the tensorflow/tensorflow repo and
# the openxla/xla repo, also include debug info for files under xla/.
build:dbg --per_file_copt=+.*,-tensorflow.*,-xla.*@-g0
build:dbg --per_file_copt=+tensorflow/core/kernels.*@-g0
# for now, disable arm_neon. see: https://github.com/tensorflow/tensorflow/issues/33360