isThreadScheduled = false;
              stillRunning = false;
              break;
            }
          }

          // Always run while _not_ holding the lock, to avoid deadlocks.
          try {
            nextToRun.call(listener);
          } catch (Exception e) { // sneaky checked exception
            // Log it and keep going.
            logger
                .get()

            // for proper cleanup, so we don't return early here
            Clock global = getClock(GLOBAL);
            global.start();
            lock.readLock().lock();
            try {
                // Process build steps in a logical order:
                // 1. Find steps that are not yet started (CREATED status)

* Fix race condition introduced by graceful termination which can lead to a deadlock in kube-proxy ([#72361](https://github.com/kubernetes/kubernetes/pull/72361), [@lbernail](https://github.com/lbernail))

            mojos.put(Thread.currentThread(), mojoDescriptor);
            boolean aggregator = mojoDescriptor.isAggregator();
            acquiredAggregatorLock = aggregator ? aggregatorLock.writeLock() : aggregatorLock.readLock();
            acquiredProjectLock = getProjectLock(session);
            if (!acquiredAggregatorLock.tryLock()) {
                Thread owner = aggregatorLock.getOwner();

* Enable audit log in local cluster ([#42379](https://github.com/kubernetes/kubernetes/pull/42379), [@xilabao](https://github.com/xilabao))
* Fix a deadlock in kubeadm master initialization. ([#43835](https://github.com/kubernetes/kubernetes/pull/43835), [@mikedanese](https://github.com/mikedanese))

If the remote handler is blocked on a send, and the client is trying to send without the remote receiving,
the operation would become deadlocked if the channels are full.

### Typed handlers

Typed handlers are handlers that have a specific type for the request and response payloads.

```go
    // Create a typed handler.

	// - missing format.json (unformatted drive)
	// - format.json is valid but invalid 'uuid'
	if err = p.checkDiskStale(); err != nil {
		return ctx, done, err
	}

	// Disallow recursive tracking to avoid deadlocks.
	if ctx.Value(healthDiskCtxKey{}) != nil {
		done = p.updateStorageMetrics(s, paths...)
		return ctx, done, nil
	}

	if contextCanceled(ctx) {
		return ctx, done, ctx.Err()
	}

completely eliminate the IP allocation issue. More information [here](https://github.com/kubernetes/kubernetes/issues/21523#issuecomment-191498969)
  * Daemon becomes unresponsive (rarely) due to kernel deadlocks. More information [here](https://github.com/kubernetes/kubernetes/issues/21866#issuecomment-189492391)

### Provider-specific Notes

#### Various

   Core changes:

  * Change default clusterCIDRs from /16 to /14 allowing 1000 Node clusters by default.
  * Allow Debian Jessie on GCE.
  * Node problem detector addon pod detects and reports kernel deadlocks.
* OpenStack
  * Provider added.
* VSphere:
  * Provider updated.

## Previous Releases Included in v1.3.0

- [v1.3.0-beta.3](CHANGELOG.md#v130-beta3)
- [v1.3.0-beta.2](CHANGELOG.md#v130-beta2)

*   Fixes a heap buffer overflow in `Transpose`
    ([CVE-2021-41216](https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2021-41216))
*   Prevents deadlocks arising from mutually recursive `tf.function` objects
    ([CVE-2021-41213](https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2021-41213))
*   Fixes a null pointer exception in `DeserializeSparse`