RE: iSCSI UNH Plugfest

[Black_David@emc.com]

> > The iSCSI initiator driver should clean up the CDB as needed (your first
> > suggestion) and keep the wire protocol clean.  If it's in an OS
> > generating old CDBs, it knows that and is best suited to fix them.
> > Don't burden the iSCSI targets with any of this.
> 
> I don't agree that the iSCSI initiator should touch the CDB. This is a ULP
> issue and the issuer of the CDB should use or not use the 
> bits as desired.

This seems to be mostly about where to locate the code that sets the
iSCSI LUN when a SCSI-2 stack sits on top of iSCSI.  Setting the iSCSI
LUN correctly is a MUST, otherwise SCSI task management will misbehave.

If one wants to split hairs, SCSI-2 code should figure out the LUN and
pass it to iSCSI code that puts it in the correct place in the iSCSI PDU.
In practice, I'm not sure this distinction is important, as I can easily
imagine this sort of design boundary being optimized out in various ways
by implementations.

On the more generic issue of SCSI-2 over iSCSI, if someone (or a set
of someones) wanted to write up an informational draft on the set of
issues that arise in doing this, that would be welcome.

Thanks,
--David

---------------------------------------------------
David L. Black, Senior Technologist
EMC Corporation, 42 South St., Hopkinton, MA  01748
+1 (508) 249-6449 *NEW*      FAX: +1 (508) 497-8500
black_david@emc.com         Cell: +1 (978) 394-7754
---------------------------------------------------