* [Breaking change, client-go]: The WaitFor function returns, probably an ErrWaitTimeout, when the done channel is closed, even if the `WaitFunc` doesn't handle the done channel. ([#72364](https://github.com/kubernetes/kubernetes/pull/72364), [@kdada](https://github.com/kdada))

* `tf.lite`
    * `Dequantize` op supports `TensorType_INT4`.
        * This change includes per-channel dequantization.
    * Add support for `stablehlo.composite`.
    * `EmbeddingLookup` op supports per-channel quantization and `TensorType_INT4` values.
    * `FullyConnected` op supports `TensorType_INT16` activation and `TensorType_Int4` weight per-channel quantization.

		return
	}
	setEventStreamHeaders(w)

	// Trace Publisher and peer-trace-client uses nonblocking send and hence does not wait for slow receivers.
	// Keep 100k buffered channel.
	// If receiver cannot keep up with that we drop events.
	traceCh := make(chan []byte, 100000)
	peers, _ := newPeerRestClients(globalEndpoints)

	}

	return dfi.ToObjectInfo(bucket, object, opts.Versioned || opts.VersionSuspended), nil
}

// Send the successful but partial upload/delete, however ignore
// if the channel is blocked by other items.
func (er erasureObjects) addPartial(bucket, object, versionID string) {
	globalMRFState.addPartialOp(PartialOperation{
		Bucket:    bucket,
		Object:    object,
		VersionID: versionID,

	}()
	go func() {
		defer close(errCh)
		// Merge all entries from all disks.
		// We leave quorum at 1, since entries are already resolved to have the desired quorum.
		// mergeEntryChannels will close 'merged' channel upon completion or cancellation.
		errCh <- mergeEntryChannels(ctx, entries, merged, 1)
	}()

	return nil
}

// HealObjectFn closure function heals the object.

	wg.Add(2)
	go publish(ReportMetrics(GlobalContext, c.metricsGroups))
	go publish(globalNotificationSys.GetClusterMetrics(GlobalContext))
	wg.Wait()
}

// ReportMetrics reports serialized metrics to the channel passed for the metrics generated.
func ReportMetrics(ctx context.Context, metricsGroups []*MetricsGroupV2) <-chan MetricV2 {
	ch := make(chan MetricV2)
	go func() {
		defer xioutil.SafeClose(ch)

	}
	diskHealthCheckOK(ctx, nil)
	return parts, nil
}

// ReadMultiple will read multiple files and send each back as response.
// Files are read and returned in the given order.
// The resp channel is closed before the call returns.
// Only a canceled context will return an error.
func (s *xlStorage) ReadMultiple(ctx context.Context, req ReadMultipleReq, resp chan<- ReadMultipleResp) error {
	defer xioutil.SafeClose(resp)

- Fixed a storage bug around multipath. iSCSI and Fibre Channel devices attached to nodes via multipath now resolve correctly if partitioned. ([#129183](https://github.com/kubernetes/kubernetes/pull/129183), [@RomanBednar](https://github.com/RomanBednar)) [SIG Storage]

* [stable] Support Attach and Detach operations for ReadWriteOnce (RWO) volumes that use iSCSI and Fibre Channel plugins.

* [stable] Expose storage usage metrics

  * The available capacity of a given Persistent Volume (PV) is available by calling the Kubernetes metrics API.

* [stable] Volume plugin metrics

### Bug or Regression

- Fixed a storage bug around multipath. iSCSI and Fibre Channel devices attached to nodes via multipath now resolve correctly if partitioned. ([#129180](https://github.com/kubernetes/kubernetes/pull/129180), [@RomanBednar](https://github.com/RomanBednar)) [SIG Storage]