Ephemeral memory volumes have an option to specify the size, but this doesn't work with OSC (kata)

example snippet

 volumes:
    - name: container-storage
      emptyDir:
        medium: Memory
        sizeLimit: 1Gi

Steps to reproduce

1. Create a Kata pod using sizeLimit
2. Exec a shell in the pod and check the size of the volumemount to see if it matches with the sizeLimit
3.

Expected result

The size of the volumemount should match the sizeLimit. 

Actual result

The size of the volumemount should match the sizeLimit. But it uses the default - 50% of total memory config.

Impact

Customers using build pipeline on OSC will not be able to set the size of the container storage (eg. /var/lib/containers)

It won't be able to set the size limit for ephemeral volumes and have to rely on workarounds.

Env

<Where was the bug found, i.e. OCP build, operator build, kata-containers build, cluster infra, test case id>

OCP 4.13 when working on https://issues.redhat.com/browse/KATA-2229

Additional helpful info

<logs, screenshot, doc links, etc.>

Reproducer

apiVersion: v1
kind: Pod
metadata:
  name: buildah-kata
  annotations:
    io.katacontainers.config.hypervisor.default_memory: "6144"
    io.katacontainers.config.hypervisor.default_vcpus: "2"
spec:
  runtimeClassName: kata
  containers:
    - name: buildah
      image: quay.io/buildah/stable:v1.30
      command: ["sh", "-c"]
      args:
      - sleep infinity
      securityContext:
        privileged: true
      volumeMounts:
        - name: container-storage
          mountPath: /var/lib/containers
  volumes:
    - name: container-storage
      emptyDir:
        medium: Memory
        sizeLimit: 1Gi

After deploying the pod, exec a shell and check the size of `/var/lib/containers`

It'll be approx 3Gi (50% of the VM memory) and not 1Gi as requested via sizeLimit.

This works fine for runc pods