RE: iscsi Rev 7.93 : text cmd & bookmarks.

["Julian Satran" <Julian_Satran@il.ibm.com>]

Targets will have an indexing mechanism based on TTT and the bookmark will
reuse it.
The 1 outstanding text command is a resource statement.

Julo

"Ayman Ghanem" <aghanem@cisco.com>@ece.cmu.edu on 17-09-2001 22:55:38

Please respond to "Ayman Ghanem" <aghanem@cisco.com>

Sent by:  owner-ips@ece.cmu.edu


To:   <ips@ece.cmu.edu>
cc:
Subject:  RE: iscsi Rev 7.93 : text cmd & bookmarks.



I could be missing something, but I am not sure why do we have TTT in text
cmd/rsp if there MUST be only one outstanding text cmd per connection. Text
command lookup implies to me more than one outstanding cmd, but that would
be a protocol error.

-Ayman

> -----Original Message-----
> From: owner-ips@ece.cmu.edu [mailto:owner-ips@ece.cmu.edu]On Behalf Of
> Santosh Rao
> Sent: Monday, September 17, 2001 1:26 PM
> To: IPS Reflector
> Subject: iscsi Rev 7.93 : text cmd & bookmarks.
>
>
> Hi Julian,
>
> Thanks for adding the target transfer tag field to the text command
> & text response. This allows for simpler indexing for targets without
> having to special case the text message lookups.
>
> Also, as you've observed and incorporated, the target transfer tag field
> itself serves as a bookmark and so, the bookmark field has been removed.
>
> One question remains however. Since the target transfer tag field serves
> as a bookmark and a TTT of 0xFFFFFFFF serves as an indication that this
> is a new text command (i.e. no bookmark present), of what value add is
> the bookmark bit ? Why not remove the bookmark bit itself and only use
> the TTT ?
>
> Keeping both the bookmark bit and the TTT is providing 2 ways of
> indicating a new text cmd. I would be in favor of removal of the
> bookmark bit.
>
> Regards,
> Santosh
>