Re: iSCSI : On the subject of R2T and Task Tags.

[Pierre Labat <pierre_labat@hp.com>]

Matt Wakeley wrote:

> The task tags in iSCSI should work the same as they do in FCP:
>
> The initiator supplies a initiator task tag per task that it uses to associate
> any received messages from the target to that task. [a "undefined" initiator
> task tag should have a value of 0xffffffff since a value of zero is a valid
> number for a table look up ].
>
> The target supplies a target task tag per task that it sends to the initiator
> where a response from the initiator is expected.  The target uses the target
> task tag to associate any solicited received messages from the initiator
> (other than the initial command) to that task.
>
> For SCSI READ commands, there is no need to specifiy a target task tag.
> Typically in FCP today, the targets to NOT assign target task tags (RX_IDs in
> FC terms) for READs.
>
> For SCSI WRITE commands, the target MAY assign a target task tag to the I/O
> that it uses when sending R2Ts to the initiator.  The initiator must then echo
> that target task tag in replies it sends to the target.  I strongly object to
> the words in section 2.16 "All outstanding R2T should have different Target
> Transfer Tags".  This specification should not dictate to the target how it
> chooses or uses the Target Task Tags.

Matt,

I agree that the draft must not impose to the target to have a different tag
for each R2T (it is up to the target implementation to decide).  I disagree
to impose a unique target tag for the whole IO.  It imposes the target to do
score boarding
to know when all the data have been received. Letting the possiblility to the
target
to have one tag per R2T (or add a subtag identifying the R2T as proposed Santosh)

allows the target to dispense with the score boarding. The target knows that all
the
data has been received when a data PDU with the "F" bit set has been received
for each R2T.

Regards,

Pierre


>
>
> -Matt Wakeley
> Agilent Technologies