LOCK_SH                         = 0x1
	LOCK_EX                         = 0x2
	LOCK_NB                         = 0x4
	LOCK_UN                         = 0x8
	POLLIN                          = 0x0003
	POLLOUT                         = 0x0004
	POLLPRI                         = 0x0010
	POLLERR                         = 0x0020
	POLLHUP                         = 0x0040
	POLLNVAL                        = 0x0080

	// ensure types are installed
	_ "k8s.io/kubernetes/pkg/apis/core/install"
)

// TestWorkloadDefaults detects changes to defaults within PodTemplateSpec.
// Defaulting changes within this type can cause spurious rollouts of workloads on API server update.
func TestWorkloadDefaults(t *testing.T) {
	t.Run("enabled_features", func(t *testing.T) { testWorkloadDefaults(t, true) })

        TypeID::get<TF::ResizeNearestNeighborGradOp>(),
        TypeID::get<TF::ReverseSequenceOp>(),
        TypeID::get<TF::RightShiftOp>(),
        TypeID::get<TF::RintOp>(),
        TypeID::get<TF::RollOp>(),
        TypeID::get<TF::RoundOp>(),
        TypeID::get<TF::SegmentSumV2Op>(),
        TypeID::get<TF::SegmentProdV2Op>(),
        TypeID::get<TF::SegmentMinV2Op>(),
        TypeID::get<TF::SegmentMaxV2Op>(),

	// registry that knows how to measure usage for objects
	registry quota.Registry

	// TODO these are used together to bucket items by namespace and then batch them up for processing.
	// The technique is valuable for rollup activities to avoid fanout and reduce resource contention.
	// We could move this into a library if another component needed it.
	// queue is indexed by namespace, so that we bundle up on a per-namespace basis

      : RewritePattern(
            RollOp::getOperationName(), 1, context,
            {ConstOp::getOperationName(), SliceOp::getOperationName(),
             ConcatV2Op::getOperationName()}) {}

  LogicalResult matchAndRewrite(Operation *op,
                                PatternRewriter &rewriter) const override {
    auto tf_roll_op = cast<RollOp>(op);

    auto input_ty =

    "denylist" where possible. Please see
    [this list](https://developers.google.com/style/word-list#blacklist) for
    more context.
*   Adds `tf.config.experimental.mlir_bridge_rollout` which will help us rollout
    the new MLIR TPU bridge.
*   Adds `tf.experimental.register_filesystem_plugin` to load modular filesystem
    plugins from Python

## Thanks to our Contributors

	EPOLLHUP                                    = 0x10
	EPOLLIN                                     = 0x1
	EPOLLMSG                                    = 0x400
	EPOLLONESHOT                                = 0x40000000
	EPOLLOUT                                    = 0x4
	EPOLLPRI                                    = 0x2
	EPOLLRDBAND                                 = 0x80
	EPOLLRDHUP                                  = 0x2000

	RESOLVE_NO_SYMLINKS   = 0x4
	RESOLVE_NO_XDEV       = 0x1
)

type PollFd struct {
	Fd      int32
	Events  int16
	Revents int16
}

const (
	POLLIN   = 0x1
	POLLPRI  = 0x2
	POLLOUT  = 0x4
	POLLERR  = 0x8
	POLLHUP  = 0x10
	POLLNVAL = 0x20
)

type sigset_argpack struct {
	ss    *Sigset_t
	ssLen uintptr
}

type SignalfdSiginfo struct {
	Signo     uint32

    return success();
  }
};

/// Converts a TF::RollOp to HLO. Only support 0D axis and shift case, and axis
/// have to be a constant.
class ConvertRollOp : public OpRewritePattern<TF::RollOp> {
 public:
  using OpRewritePattern::OpRewritePattern;
  LogicalResult matchAndRewrite(TF::RollOp op,
                                PatternRewriter &rewriter) const override {

// generated by the class object during its lifetime.
//
// In theory, this class is not necessary because we should simply take
// the TF function name and use it as MLIR function name. However, for some
// unknown reasons (callout for investigation in b/142268695), keeping the
// function names unchanged in an MLIR roundtrip causes test failures.
// TODO(b/142268695) Re-evaluate whether we need this class v.s. directly using