volume creation fails even after having enough capacity on node

[iyashu]

Even we have enough capacity on node (correctly published under csi storage capacity resource), pvc creation is failing with following error -

2021-08-26T01:25:12.850848431+05:30 stderr F E0826 01:25:12.850660       1 device-util.go:161] findBestPart Failed
2021-08-26T01:25:12.850852115+05:30 stderr F E0826 01:25:12.850687       1 volume.go:243] error syncing 'openebs-device/pvc-9c9d05b5-3dc1-4201-84bb-6f416532e508': free space of 9060352MiB not found on disk name: xxxx-*, requeuing

Capacity asked by PVC -

apiVersion: v1
kind: PersistentVolumeClaim
metadata:
  annotations:
    device.csi.openebs.io/csi-volume-name: pvc-9c9d05b5-3dc1-4201-84bb-6f416532e508
    volume.beta.kubernetes.io/storage-provisioner: device.csi.openebs.io
    volume.kubernetes.io/selected-node: node-xyz
  creationTimestamp: "2021-08-25T18:56:08Z"
  name: ...
  namespace: ...
  resourceVersion: "105909663"
  uid: 9c9d05b5-3dc1-4201-84bb-6f416532e508
spec:
  accessModes:
  - ReadWriteOnce
  resources:
    requests:
      storage: 9500G
  storageClassName: hdd-xxxx-part-ext4-0
  volumeMode: Filesystem
status:
  phase: Pending

Relevant CSI storage capacity resource -

apiVersion: storage.k8s.io/v1alpha1
capacity: 9060351Mi
kind: CSIStorageCapacity
...

Issue is happening due to conversion of pv size to nearest Gi unit here. Any specific reason of doing so? Can we remove the Gi unit conversion?