Add simple GPU support and improve rclone secret usage in Helm chart

[rdmcguire]

This PR primarily provides a simple way for enabling support for an Nvidia GPU in Kubernetes. Additionally, it copies all keys from the rclone secret into /etc/rclone, which is useful if you want to ship credentials or other files, such as a private key for rclone via sftp.

The values.yaml interface for enabling GPU support is in these keys:

pms:
  # Enabling this will add nvidia.com/gpu: 1 to limits, and will set
  # environment for the nvidia operator
  gpu:
    nvidia:
      enabled: false

Additionally the config init container command and args are modified to copy all secret keys into the config-data mount, and more useful extraVolumes / extraVolumeMounts comments were added for /dev/dri specifically as it is likely to be common for plex.

[MarshallAsch]

Nice addition.

We have 2 requests, could you please add a minor version bump the the Chart.yaml file, and re-generate the values file documentation in the README using frigate. (there is a link at the bottom of the readme that can be used)

[cilindrox]

LGTM - thanks @rdmcguire

Deferring to @MarshallAsch for final approval.

[cilindrox]

new version shipped. Thanks @rdmcguire !