runtime-rs: Enhancing Direct Volume Mount Handling with Patching Support for current infra(K8S/CSI)

[Apokleos]

Which feature do you think can be improved?
The current infrastructure for Kata containers, which includes Kubernetes (K8s), Container Storage Interface (CSI), Container Runtime Interface (CRI), and Containerd, lacks the capability to effectively manage direct volumes. This limitation necessitates workarounds, such as modifying mount type and options after the volume's creation.

Direct volumes, unlike other types of storage volumes, are directly attached to the virtual machine (VM) hosting the Kata container. This direct attachment necessitates additional information, such as the block device path and filesystem type, for proper mounting within the container.

However, the current infrastructure fails to convey this crucial information during the volume creation process. As a result, workarounds are required to rectify the mount configuration after the volume has been created. This involves adjusting the mount type to reflect the direct volume's nature and specifying the appropriate mount options to ensure proper functionality within the Kata container environment.

The need for these workarounds highlights a gap in the current infrastructure's ability to seamlessly handle direct volumes for Kata containers. Addressing this limitation would eliminate the need for post-creation modifications and streamline the overall volume management process for Kata containers.

One word：The current infra(K8S, CSI, CRI, Containerd) for Kata containers is unable to properly handle direct volumes, resulting in the need for workarounds like patching up mount type and options after it created.

How can it be improved?

Do some work to patch up mount type and options after direct volume created during amend OCI spec.