Support for pre-provisioned volumes

[WanzenBug]

In theory, we should be able to support pre-provisioned volumes like this:

1. Create a LINSTOR volume:

   linstor rd c pv1
   linstor vd c pv1 1g
   linstor rd ap pv1
   

2. Create a persistent volume and claim:

   apiVersion: v1
   kind: PersistentVolume
   metadata:
     name: pv1
   spec:
     accessModes:
     - ReadWriteOnce
     capacity:
       storage: 1Gi
     csi:
       driver: linstor.csi.linbit.com
       fsType: ext4
       volumeAttributes: {}
       volumeHandle: pv1
     persistentVolumeReclaimPolicy: Delete
     volumeMode: Filesystem
   ----
   apiVersion: v1
   kind: PersistentVolumeClaim
   metadata:
     name: pv1
   spec:
     storageClassName: "" # Empty string must be explicitly set otherwise default StorageClass will be set
     volumeName: pv1
     accessModes:
     - ReadWriteOnce
     resources:
       requests:
         storage: 1Gi

When using the pod we get the following issue:

  Warning  FailedAttachVolume  85s (x4 over 11m)    attachdetach-controller  (combined from similar events): AttachVolume.Attach failed for volume "pv1" : rpc error: code = Internal desc = ControllerPublishVolume failed for mwa1: unable to find CSI volume annotation on resource {Name:pv1 ExternalName: Props:map[DrbdOptions/Resource/on-no-quorum:io-error DrbdOptions/Resource/quorum:majority DrbdOptions/auto-verify-alg:crct10dif-generic DrbdPrimarySetOn:DEMO1.LINSTOR-DAYS.AT.LINBIT.COM] Flags:[] LayerData:[{Type:DRBD Data:0xc00003b040} {Type:STORAGE Data:<nil>}] Uuid:2d154ba7-bd20-4729-9ddd-f480eb6b301a ResourceGroupName:DfltRscGrp}

This points in a similar direction as #60

I believe removing the need for the magic annotation would fix this issue.