FC Reservations

Binford, Charles charles.binford at symbios.com
Mon Mar 30 13:57:00 PST 1998


* From the T10 (formerly SCSI) Reflector (t10 at symbios.com), posted by:
* "Binford, Charles" <charles.binford at symbios.com>
*
Brian, this question was asked last fall on the disk_attach reflector.
At that time I asked Dave Peterson as the editor of PLDA-2 to add
reservation into table 16.  Between then and now PLDA-2 as turned into
FC-TAPE, so I'm not sure if the answer to this question has a document
home or not.

Anyway, my understanding of things gives the answers below.

Charles Binford
Symbios Inc.



 ----------
>From: bmckean at mylexboulder.com
>To: SCSI Reflector; Fibre Channel Reflector
>Subject: FC Reservations
>Date: Monday, March 30, 1998 11:53AM
>
>
>I'd like some help to understand when reservations (using
>Reserve/Release, not persistent reservations) are cleared under FC-AL.
>
>1. Which loop operations cause ALL device reservations to be cleared
and
>which cause only reservations cleared for the initiator that performs
>the operation?
>
>     FC-PLDA rev 2.1 table 16 lists when to clear "FCP SCSI Target
>     Objects"  but does not list reservations.
>     1.1 The easy ones for clear ALL would be:
>          a.  Device power on or hard reset
>          b. Target Reset task management function ( FCP_CNTL  Target
>          Reset)

agreed

>
>     1.2 What about LIP reset L_PORT ?

If I'm the device being selectively reset, then yes, clear the
reservations (treat as hard reset).  If someone else is being reset, or
any other flavor of LIP, then no, don't clear the reservation.

>
>     1.2 What about PLOGI ?

Yes, clear the reservation.  This includes explicit and implicit LOGO as
well.  Note, that not "checking in" after a normal LIP can cause an
implicit logout, thus clear a reservation.

>
>     1.3 What about PRLI ?
>          FCP r12 section 6.2.5 New or repeated PRLI (second paragraph)
>          mentions reservations:

Yes, clear the reservation.

>
>               Immediately after the execution of a PRLI, both members
>               of the
>               image pair shall have the same state as they
>               would have after a hard reset or a power on with respect
>               to each other.
>          and
>
>               Tasks, reservations, status, and MODE SELECT
>               parameters for
>               other initiators are not affected.

I think you misread FCP.  Here is the exact text I think you are
referring to:
"Immediately after the execution of a PRLI, both members of the image
pair shall have the same state as they
would have after a hard reset or a power on with respect to each other.
No tasks, reservations or status except
those previously established between other image pairs sharing one of
the process associators shall be present
in either SCSI device."

The confusing part is in the last sentence.  The clause "...except those
previously established between other image pairs sharing of the process
associators...." refers to those using process associators.  For all of
us NOT using process associators I would change the last sentence to:
"No tasks, reservations or status shall be present in either SCSI
device."

>
>           This seems contradictory.  After hard reset or power on ALL
>          reservations would be cleared.   How can the target behave as
>          if power on occurred but still preserve reservations for
other
>          initiators?
>
>     Any others I didn't think of?
>
>
>2. A separate question:   PLOGI in table 16 it says to restore mode
>pages from saved pages only for port initiating action.  It shows for
>all other SCSI initiators that mode pages are NOT restored from saved.
>Does this imply that mode pages must be kept on a per initiator basis?
>
>
>Thanks,
>
>Brian McKean
>
> --
>
>Brian McKean                             Voice: (303)381-4246
>Mylex Corporation External Products      Fax:   (303)413-0464
>4900 Pearl East Circle Suite 104         brianm at mylexboulder.com
>Boulder, CO 80301-6108
>http://www.mylex.com
>
>
*
* For T10 Reflector information, send a message with
* 'info t10' (no quotes) in the message body to majordomo at symbios.com





More information about the T10 mailing list