- The feature gate was mentioned as `csiMigrationRBD` where it should have been `CSIMigrationRBD` to be in parity with other migration plugins. This release correct the same and keep it as `CSIMigrationRBD`.
  
  users who have configured this feature gate as `csiMigrationRBD` has to reconfigure the same to `CSIMigrationRBD` from this release. ([#107554](https://github.com/kubernetes/kubernetes/pull/107554), [@humblec](https://github.com/humblec)) [SIG Storage]

- The feature gate was mentioned as `csiMigrationRBD` where it should have been `CSIMigrationRBD` to be in parity with other migration plugins. This release correct the same and keep it as `CSIMigrationRBD`.
    users who have configured this feature gate as `csiMigrationRBD` has to reconfigure the same to `CSIMigrationRBD` from this release. ([#107554](https://github.com/kubernetes/kubernetes/pull/107554), [@humblec](https://github.com/humblec))

* Add masquerading rules by default to GCE/GKE ([#55178](https://github.com/kubernetes/kubernetes/pull/55178), [@dnardo](https://github.com/dnardo))
* kubeadm now produces error during preflight checks if swap is enabled. Users, who can setup kubelet to run in unsupported environment with enabled swap, will be able to skip that preflight check. ([#55399](https://github.com/kubernetes/kubernetes/pull/55399), [@kad](https://github.com/kad))

when asking for "All" devices on a node, Kubernetes <= 1.32 proceeded to schedule pods onto nodes with no devices by not allocating any devices for those pods. Kubernetes 1.33 changes that to only picking nodes which have at least one device. Users who want the "proceed with scheduling also without devices" semantic can use the upcoming prioritized list feature with one sub-request for "all" devices and a second alternative with "count: 0". ([#129560](https://github.com/kubernetes/kubernetes/pull/129560),...

- Changed how the API server handles updates to `.spec.defaultBackend` of Ingress objects.

- Drop additional rule requirement (cronjobs/finalizers) in the roles who use kubectl create cronjobs to be backwards compatible ([#124883](https://github.com/kubernetes/kubernetes/pull/124883), [@ardaguclu](https://github.com/ardaguclu)) [SIG CLI]
- Fix a bug that Pods could stuck in the unschedulable pod pool 

call site.
</p>

<p>
Given the function and calls
</p>
<pre>
func Greeting(prefix string, who ...string)
Greeting("nobody")
Greeting("hello:", "Joe", "Anna", "Eileen")
</pre>

<p>
within <code>Greeting</code>, <code>who</code> will have the value
<code>nil</code> in the first call, and
<code>[]string{"Joe", "Anna", "Eileen"}</code> in the second.
</p>

Other then that, this release contains no other changes from 1.5.4.

The vulnerability is tracked in http://issue.k8s.io/43459.

**Who is affected?**

Only Kubernetes 1.5.0-1.5.4 installations that do all of the following:
* Enable the PodSecurityPolicy API (which is not enabled by default):
  * `--runtime-config=extensions/v1beta1/podsecuritypolicy=true`

discovery now includes the singularName used to refer to the resource. ([#43312](https://github.com/kubernetes/kubernetes/pull/43312), [@deads2k](https://github.com/deads2k))

* Enhance the garbage collection admission plugin so that a user who doesn't have delete permission of the owning object cannot modify the blockOwnerDeletion field of existing ownerReferences, or add new ownerReferences with blockOwnerDeletion=true ([#43876](https://github.com/kubernetes/kubernetes/pull/43876), [@ca...

For stable features, the `kubectl` tool follows the Kubernetes [version skew policy](https://kubernetes.io/releases/version-skew-policy/);
however, kubectl v1.21 and older do not support the new API for ephemeral containers.