QErr, Basic Queuing and SBP-2

PJohansson at aol.com PJohansson at aol.com
Thu Feb 20 10:03:15 PST 1997


* From the SCSI Reflector (scsi at symbios.com), posted by:
* PJohansson at aol.com
*
The SBP-2 ad hoc working group has embraced the concept of Basic queuing for
all of the reasons cited by Jim McGrath and others, chief of which is that
the cost of the drive can be reduced because of a simpler test and
qualification matrix.

However, SBP-2 is intended to be useful for multi-initiator devices as well
as low-end, single-initiator devices. In fact, SBP-2 is structured so that
there is far less code complexity to abort an initiator's task set (a linked
list of commands, in SBP-2 parlance) than there is to abort all initiators'
task sets.

Because of this, the working group supports 96-277r1 and the proposal to
modify the SAM-2 Basic queuing model to reference QErr == 11b (the affected
initiator's task set is aborted subsequent to a CHECK CONDITION).

I also think this causes no problem to a single-initiator device, since there
is only one task set. It seems to me that the diffficulty, if any, is for the
device to refuse commands from more than one initiator at a time. In SBP-2
this is not a problem: an initiator must login before it can send commands
and a single-initiator SBP-2 device accepts only one login at a time. One
option for a parallel SCSI device would be to return BUSY when a second,
third, etc. initiator attempted to send a command while a non-empty task set
existed for another initiator.

Regards,

Peter Johansson

Congruent Software, Inc.
3998 Whittle Avenue
Oakland, CA  94602

(510) 531-5472
(510) 531-2942 FAX

pjohansson at aol.com

*
* For SCSI Reflector information, send a message with
* 'info scsi' (no quotes) in the message body to majordomo at symbios.com




More information about the T10 mailing list