Handle incorrectly formatted disks

SAS disks may be formatted with an increased sector size and extra protection information (DIF). With support from the HBA, this means checksums will be computed/verified by the HBA for every block. This is redundant and a waste of bus bandwidth with ZFS. These disks should be reformatted to use a normal sector size without protection information before a pool can be created.

At the least, we should alert the user about these disks and prevent them from creating a pool. Even better would be to do the reformatting for them if needed, though the time-consuming nature of this process must be communicated properly.

Another complication is that FreeBSD lacks support for DIF and ignored this format error in some cases. So, users coming from CORE to SCALE may have had incorrectly formatted disks when the pool was created in CORE and then in SCALE with DIF supported the disks are now unreadable. These users need to disable the protection capability, either at the driver level if possible (i.e., mpt3sas.prot_mask=0 kernel option) or at the scsi_disk level for each affected disk (i.e., `echo 0 > /sys/class/scsi_disk/H:C:I:L/protection_type`). Not all drivers support disabling protection at the driver level. It would be slick if this could be configured through the UI as an alternative to reformatting, so that existing pools are able to be imported.