|
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index] Re: iSCSI: Indication of final data-out sequenceThere is no need for an additional bit. Every outstanding R2T has a target task tag (and a number...) The target has to keep track of what he got and send the status after the last R2T sequence. Thee is also an ordering delivery requirement but that is weaker. Julo Rishabh Srivastav <rishabh_srivastav@yahoo.com> on 05-07-2001 12:18:47 Please respond to Rishabh Srivastav <rishabh_srivastav@yahoo.com> To: Julian Satran/Haifa/IBM@IBMIL, ips@ece.cmu.edu cc: Subject: Re: iSCSI: Indication of final data-out sequence Julo, I'm still not convinced. There can be multiple data-out PDU's in response to one R2T. In this case the F bit indicates the end of sequence corresponding to that particular R2T and not for the whole write operation. However, the last_sequence bit in FC-FS is something which indicates the final data IU for that whole exchange. So, how can 'F' bit be equated to the last_sequence bit ?? The semantics of 'F' bit is same as last_sequence bit in the case of Data-In PDUs and not in the case of data-out PDU. Where am i going wrong in my assumptions ? Please clarify. Thanks, Rishabh --- julian_satran@il.ibm.com wrote: > > > the F bit. Julo > > A doubt regarding the final sequence of data-out > PDUs > : My apologies if this had already been discussed > long > time ago. Couldn't locate it in the huge mail > repository. > In FC, the last sequence of an exchange is indicated > by a 'Last_Sequence' bit in F_CTL of FC-FS header. > What is the equivalent of it in iSCSI, specifically > in > the case of Data-Out ? If there is none, will not > the > presence of a similar bit be useful for iSCSI-FCP > gateways ? > > Thanks, > Rishabh > __________________________________________________ Do You Yahoo!? Get personalized email addresses from Yahoo! Mail http://personal.mail.yahoo.com/
Home Last updated: Tue Sep 04 01:04:21 2001 6315 messages in chronological order |