iSCSI: No-renegotiation rule inadequately described

[Martins Krikis <mkrikis@yahoo.com>]

Before I start, a philosophical question:

What is the point of allowing Reject-ed keys
to be renegotiated?

Now the problem. 
http://www.pdl.cmu.edu/mailinglists/ips/mail/msg09733.html
claims that that it is OK to renegotiate keys that
have been answered with the reserved value "Reject".
I don't see where this would be mentioned in the
draft. What I see is this (page 73, very top):

  Neither the initiator nor the target should
  attempt to negotiate a parameter more than
  once during login.

Similar sentence on page 79, very bottom.

If we now imagine the following scenario:

  I -> T: ImmediateData=junk1
  T -> I: ImmediateData=Reject
  I -> T: ImmediateData=junk2

Doesn't it look like Initiator is _attempting_
to negotiate this parameter again? Thus, the
spec the way I read it disallows this (and I
like it), however the above link claims otherwise.
So, which way is it?

If it is allowed, then the draft should say so,
and it should also say whether keys answered
with Irrelevant may be renegotiated. I presume
keys answered with NotUnderstood may too, although
I'm not planning to even notice such an event.

Thanks,

  Martins Krikis, Intel Corp.

Disclaimer: these opinions are mine and may not
            be those of my employer

 

__________________________________________________
Do You Yahoo!?
LAUNCH - Your Yahoo! Music Experience
http://launch.yahoo.com