In the rewritten program, the other goroutine can observe <code>2</code>, which was previously impossible.
</p>

<p>
Not introducing data races also means not assuming that loops terminate.
For example, a compiler must in general not move the accesses to <code>*p</code> or <code>*q</code>
ahead of the loop in this program:
</p>

<pre>
n := 0
for e := list; e != nil; e = e.next {
	n++
}

  
  However, the Job controller creates a replacement Pod as soon as the termination becomes apparent.
  Once the pod terminates, the Job controller evaluates `.backoffLimit` and `.podFailurePolicy` for the
  relevant Job, taking this now-terminated Pod into consideration.
  
  This behavior is limited to Jobs with `.spec.podFailurePolicy` set, and only when those two feature

* only include running and pending pods in daemonset should place calculation ([#23929](https://github.com/kubernetes/kubernetes/pull/23929), [@mikedanese](https://github.com/mikedanese))
* A pod never terminated if a container image registry was unavailable ([#23746](https://github.com/kubernetes/kubernetes/pull/23746), [@derekwaynecarr](https://github.com/derekwaynecarr))

* Fixed a bug in the single-numa-policy of the TopologyManager. Previously, best-effort pods would result in a terminated state with a TopologyAffinity error. Now they will run as expected. ([#83777](https://github.com/kubernetes/kubernetes/pull/83777), [@lmdaly](https://github.com/lmdaly))



# v1.16.4

* Never prevent deletion of resources as part of namespace lifecycle ([#48733](https://github.com/kubernetes/kubernetes/pull/48733), [@liggitt](https://github.com/liggitt))
* kubectl: Fix bug that showed terminated/evicted pods even without `--show-all`. ([#48786](https://github.com/kubernetes/kubernetes/pull/48786), [@janetkuo](https://github.com/janetkuo))

   * includes exceptional exits, so all control paths involving signalling must be protected by a
   * finally block.
   *
   * Further optimizations of this algorithm become increasingly subtle. A wait that terminates
   * without the guard being satisfied (due to timeout, but not interrupt) can then immediately exit
   * the monitor without signalling. If it timed out without being signalled, it does not need to

- Fixes regression in 1.25.10 causing running pods with devices to be terminated if kubelet is restarted ([#119707](https://github.com/kubernetes/kubernetes/pull/119707), [@ffromani](https://github.com/ffromani)) [SIG Node and Testing]

#### **Other**

*   Pod addresses will now be removed from the list of endpoints when the pod is in graceful termination. ([#54828](https://github.com/kubernetes/kubernetes/pull/54828),[ @freehan](https://github.com/freehan))

- kubeadm: ensure that Pods from the upgrade preflight check `CreateJob` are properly terminated after a timeout. ([#127333](https://github.com/kubernetes/kubernetes/pull/127333), [@yuyabee](https://github.com/yuyabee)) [SIG Cluster Lifecycle]

- Fix the regression with the slow pods termination. Before this fix pods may take an additional time to terminate - up to one minute. Reversing the change that ensured that CNI resources cleaned up when the pod is removed on API server. ([#97980](https://github.com/kubernetes/kubernetes/pull/97980), [@SergeyKanzhelev](https:...