Re: ISCSI: Unsolicited data in draft v12

[Luben Tuikov <luben@splentec.com>]

"BURBRIDGE,MATTHEW (HP-UnitedKingdom,ex2)" wrote:
> 
> Julian,
> 
> The general impression I get is that the vast majority of the time an
> initiator who negotiates to use unsolicited data will use unsolicited data.
> Rod has highlighted a valid scenario whereby the initiator may choose not to
> send unsolicited data.
> 
> However, I would like to add the following statement to the spec to
> highlight the fact the Rod's proposal be used in rare situations and is by
> no means the norm.
> 
> "If an initiator negotiates the use of unsolicited data then it SHOULD
> choose to use it. If the initiator chooses to send non-immediate unsolicited
> data ..."

I too agree strongly with this. Julian, please add the above,
as it will definitely make things clearer.

Someone mentioned this in the beginning of the thread, which
basically translates to ``the initiator will have to follow suit,
and willingly abide by what it has negotiated''.

BTW, the current wording of 2.2.4/11-94 quoted here:

  ``If the initiator chooses to send non-immediate unsolicited data
    the total amount of immediate and non-immediate unsolicited data
    MUST be either the negotiated amount or all the data if the total
    amount is less than the negotiated amount for unsolicited data.
    Otherwise, regardless of the negotiated mode, an initiator may
    choose to send only immediate data or no unsolicited data at all.''

is long and cryptic at best.

Ideas like this a better conveyed by one sentence stating what the
goal is, followed by 2-3 sentences of how the draft/standard
accomplishes it. (but surely you all know this)

The above quotation misses the goal part and the rest is long and
cryptic.

-- 
Luben