Re: Piggyback status gripe

[Stephen Schaeffer <stephens@iol.unh.edu>]

Mark,

We will indeed write this into our recommended interop procedure.

We are in the process of finishing up the interop suite, but if anyone would
like a copy please ping me, and I'll send you a copy of the current version. I
intend to release the final thing in all it's glory soon.

Stephen


"Mark S. Edwards" wrote:

> Over the last several weeks, I have found myself testing against some
> version 8 initiators that do not understand that a good status can be
> carried piggyback fashion in the DATA-IN PDU, as indicated by the 'S' bit.
>
> This makes life very difficult for a target, and takes a while to analyse
> the situation to find out what is wrong.
>
> Can I draw initiator writers (especially those in the Windows arena)
> attention to the text in 3.7 of both draft 8 and draft 9 and the equivalent
> text in 10.7 of draft 10.
>
> Initiators MUST cope with both cases of either a DATA-IN with the 'S' bit
> set, or a separate SCSI Response PDU, it's not up to the target to
> telepathically cope with which behaviour the initiator expects.  For
> performance sake, most targets will piggyback the status in the DATA-IN PDU
> and this hangs a couple of initiators.
>
> Could the participants of next weeks Plugfest actively check for this
> compliance ?
>
> Regards,
>
> Mark.

--

Stephen Schaeffer
Fibre Channel and iSCSI Consortia Manager
InterOperability Lab, University of New Hampshire

Email: stephens@iol.unh.edu
Phone: 603-862-5083
Lab Phone: 603-862-0701
URL: www.iol.unh.edu