Maybe I'm not being clear with what exactly we expect from this. Comparing to hyper-v (we have a mix hypervisor environment), we can present directly the same LUN (200 GB for example) to 02 VM's and creat a SQL Cluster. No need to present this 200 GB LUN to the hyper-v host and passtrough to the VM's, mapping is directly to the VMs virtual HBAs.
The thing to be aware of here is that ESXi does not present a virtual FC HBA to VMs with NPIV. It just maps LUNs on the SCSI layer to the VMs SCSI controller. The NPIV virtual WWN is owned by the physical host currently running the VM.
You should be able to use your clustered filesystem anyways with RDMs. Using NPIV with ESXi basically only gives you a slight management advantage. For stuff like this the actual transport technology shouldn't matter, it's the SCSI layer that's important.
You should read some of the following documents, I know they target Windows clusters (apart from the last article), but you can apply the general concepts and procedures to basically any SCSI shared-disk clustering solution:
VMware KB: Microsoft Clustering on VMware vSphere: Guidelines for supported configurations
VMware KB: Disabling simultaneous write protection provided by VMFS using the multi-writer flag