What if there is no Process Login?

David Deming ddeming at scruznet.com
Thu Dec 16 13:09:14 PST 1999


* From the T10 Reflector (t10 at t10.org), posted by:
* "David Deming" <ddeming at scruznet.com>
*
Bob,

Here is my 2 cents.

Option one.

The device simply return a Check Condition status with new ASC & ASCQ codes
stating something like "device not properly logged in".

Option two:

The device return the obsoleted "Command Terminated" status code (0x22)
that nobody used due to lack of implementation. This option seems more
simple for device driver coders because they only need to check the status
byte.

regards,

daveD

----------
> From: Bob Snively <Bob.Snively at EBay.Sun.COM>
> To: fc at network.com; t10 at t10.org
> Subject: What if there is no Process Login?
> Date: Thursday, December 16, 1999 11:03 AM
> 
> *
> * From the fc reflector, posted by:
> * Bob Snively <Bob.Snively at EBay.Sun.COM>
> *
> 
> Well, if you are talking to a disk drive, it should be a simple block
> of aluminum with a non-descript NL_Port attached.  Similarly for an
> initiator, although the block of aluminum may weigh a bit more.
> 
> However, what that really means is not covered in FCP-2.  Here is my
> FCP-2 centric answer to the question, which will go in T10/99-325r2
> unless someone has a better idea:
> 
> 	8.4  Behavior if no PRLI
> 
> 	FCP-2 does not say what to do when PRLI has not been done to a
> 	device and an FCP command is sent out to it.
> 
> 	FC-PLDA section 9.7 specifies that the responder should send
> 	back a logout and discard the command.  This is not an FC-PH
> 	compliant action, and needs to be fixed up.  
> 
> 	There are actually two cases of interest. If the devices
> 	support implicit PRLI, then the FCP command is expected to be
> 	accepted and executed normally.  If the devices do not support
> 	implicit PRLI, the target device first needs to clean up the
> 	exchange by sending back a P_RJT indicating "Login required".
> 	An additional P_RJT code of "Process Login Required" may also
> 	be needed.  Whoever is most interested, the recipient or the
> 	originator, should then perform the appropriate login and
> 	continue operation.
> 
> 	This is an FC-FS problem and not an FCP-2 problem. It seems
> 	pretty clearly implied by 14.3.3.3 of FC-FS.
> 
> 	To be proposed for FC-FS, if approved by e-mail.

*
* For T10 Reflector information, send a message with
* 'info t10' (no quotes) in the message body to majordomo at t10.org




More information about the T10 mailing list