*p /= 2
*p += i
</pre>

<p>
If <code>i</code> and <code>*p</code> start equal to 2,
the original code does <code>*p = 3</code>,
so a racing thread can read only 2 or 3 from <code>*p</code>.
The rewritten code does <code>*p = 1</code> and then <code>*p = 3</code>,
allowing a racing thread to read 1 as well.
</p>

<p>
Note that all these optimizations are permitted in C/C++ compilers:

| Feature | Components |
|---------|------------|
Telemetry | Prometheus
Telemetry | Prometheus Operator
Telemetry | Grafana
Telemetry | Kiali
Telemetry | Tracing
ThirdParty | CNI

### Namespaces

The `IstioOperatorSpec` API and underlying new Helm charts offer a lot of flexibility in which namespaces features and

      // For each listener we add a task that submits it to the executor directly for the blocking
      // get use case and another task that adds it as a listener to the future to exercise both
      // racing addListener calls and addListener calls completing after the future completes.
      final Runnable listener =
          k % 2 == 0 ? collectResultsRunnable : collectResultsTimedGetRunnable;

      // For each listener we add a task that submits it to the executor directly for the blocking
      // get use case and another task that adds it as a listener to the future to exercise both
      // racing addListener calls and addListener calls completing after the future completes.
      final Runnable listener =
          k % 2 == 0 ? collectResultsRunnable : collectResultsTimedGetRunnable;

    handler.windowUpdate(streamId, increment)
  }

  @Throws(IOException::class)
  override fun close() {
    source.close()
  }

  /**
   * Decompression of the header block occurs above the framing layer. This class lazily reads
   * continuation frames as they are needed by [Hpack.Reader.readHeaders].
   */
  internal class ContinuationSource(
    private val source: BufferedSource,
  ) : Source {

    var finished: Boolean = false,
  ) : Sink {
    /**
     * Buffer of outgoing data. This batches writes of small writes into this sink as larges frames
     * written to the outgoing connection. Batching saves the (small) framing overhead.
     */
    private val sendBuffer = Buffer()

    /** Trailers to send at the end of the stream. */
    var trailers: Headers? = null

    var closed: Boolean = false

// Allows a single op at a time to be launched without blocking.
//
// DeviceThread itself is thread-safe, in that StartExecute will block if there
// is a pending execution. Since StartExecute is equivalent to grabbing a lock,
// multiple DeviceThreads should always be accessed in the same order to avoid
// deadlocks.
class DeviceThread {
 public:
  // Starts a background thread waiting for `StartExecute`.

    /**
     * The smallest message that will be compressed. We use 1024 because smaller messages already
     * fit comfortably within a single ethernet packet (1500 bytes) even with framing overhead.
     *
     * For tests this must be big enough to realize real compression on test messages like
     * 'aaaaaaaaaa...'. Our tests check if compression was applied just by looking at the size if

    # to use for pulling any images in pods that reference this ServiceAccount.
    # For components that don't use ServiceAccounts (i.e. grafana, servicegraph, tracing)
    # ImagePullSecrets will be added to the corresponding Deployment(StatefulSet) objects.
    # Must be set for any cluster configured with private docker registry.
    imagePullSecrets: []
    # - private-registry-key

	// Handlers have no versioning, so non-compatible handler changes must result in new IDs.
	handlerTest
	handlerTest2
	handlerLast
)

// handlerPrefixes are prefixes for handler IDs used for tracing.
// If a handler is not listed here, it will be traced with "grid" prefix.
var handlerPrefixes = [handlerLast]string{
	HandlerLockLock:                    lockPrefix,
	HandlerLockRLock:                   lockPrefix,