Re: iSCSI DataSN Can equal 2**32

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

John,

I'll fix the wording - and the conflicting statement about data ack.


Julo


                                                                                                                                    
                    John                                                                                                            
                    Hufferd@IBMUS        To:     Julian Satran/Haifa/IBM@IBMIL@IBMDE                                                
                                         cc:                                                                                        
                    12/01/2001           From:   John Hufferd/San Jose/IBM@IBMUS                                                    
                    11:04 AM             Subject:     Re: iSCSI DataSN Can equal 2**32(Document link: Julian Satran - Mail)         
                                                                                                                                    
                                                                                                                                    
                                                                                                                                    
                                                                                                                                    
                                                                                                                                    
                                                                                                                                    



Julian,
I think that is what I said.  The document does not say that the largest
number is 2**32-1, it says the number of PDUs sent must be 2**32-1, and
then be less then that.  I thought I understood what is meant, but it is
not what is written.  Please see my suggestions for rewrite, highlighted in
red, below.

.
.
.
John L. Hufferd
Senior Technical Staff Member (STSM)
IBM/SSG San Jose Ca
Main Office (408) 256-0403, Tie: 276-0403,  eFax: (408) 904-4688
Home Office (408) 997-6136, Cell: (408) 499-9702
Internet address: hufferd@us.ibm.com


Julian Satran@IBMIL
11/30/2001 08:51 PM

To:   John Hufferd/San Jose/IBM@IBMUS@IBMDE
cc:
From: Julian Satran/Haifa/IBM@Haifa/IBM@IBMIL
Subject:  Re: iSCSI DataSN Can equal 2**32  (Document link: John Hufferd)

John,

The limitation to 2**31-1 stems from the ExpDataSN - the number of data/R2T
PDUs the target has sent.
The largest number that can be announced is 2**31-1 and that means 2**32
PDUs (numbering starts from 0).

Julo


                                                                                                                                    
                    John                                                                                                            
                    Hufferd@IBMUS        To:     Julian Satran/Haifa/IBM@IBMIL@IBMDE                                                
                                         cc:     ips@ece.cmu.edu                                                                    
                    11/30/2001           From:   John Hufferd/San Jose/IBM@IBMUS                                                    
                    08:46 AM             Subject:     iSCSI DataSN Can equal 2**32                                                  
                                                                                                                                    
                                                                                                                                    
                                                                                                                                    



Julian,
At  3.7.5 there is a statement that data sequence MUST contain less then
2**32-1 numbered PDUs.  I did not understand that since the count starts at
0, which should mean that 2**32 PDUs could be possible.  So I do not
understand the -1 in the count of PDUs, and I then do not understand "less
then" (which I guess would mean 2**32-2).  I think the number should be
able to go from 0 to 2**32-1, inclusive, which is 2*32 different PDUs and
not "less then 2**32-1".

Perhaps the statement should read that the sequence number of the Data PDU
can not be greater then 2**32-1.  Or the total number of Data PDUs in a
Sequence can not be greater then 2**32.

But perhaps I missed something.

.
.
.
John L. Hufferd
Senior Technical Staff Member (STSM)
IBM/SSG San Jose Ca
Main Office (408) 256-0403, Tie: 276-0403,  eFax: (408) 904-4688
Home Office (408) 997-6136, Cell: (408) 499-9702
Internet address: hufferd@us.ibm.com