Persistent Reserve Document, X3T10/95-229r0
Gene Milligan
Gene_Milligan at notes.seagate.com
Tue May 30 05:39:23 PDT 1995
Bob Snively wrote:
>7) Target Reset
>As presently defined in SAM, TARGET RESET and CLEAR TASK SET clear all tasks
for all
>initiators. This proposal retains that property and includes that definition
in the proposed
>modifications to clause 5.4 of SPC. Some proposals have suggested that TARGET
RESET
>should only reset the tasks for the initiator that generated the task
management function.
>Unless the committee directs otherwise, I propose that we leave the
definitions unchanged and
>accept the clear understanding that TARGET RESET will mess up all initiators
and should not
>be used. At least it will not clear persistent reservations.
The proposal is crafted to allow a mixture of legacy initiators (don't speak
persistent reservations) and Persistent Reservation initiators. What is the
benefit of having persistent reservations withstanding TARGET RESET if all the
reserved tasks are blown away? It seems to me you can postulate a useful world
of throwing away (or retrofitting) all initiators that don't speak persistent
reservations if the system needs to practice safe persistent reservations. The
net result could be a cleaner persistent reservation world.
Gene
More information about the T10
mailing list