STATE_CLEAR and STATE_SET

PJohansson at aol.com PJohansson at aol.com
Wed Jun 4 23:05:51 PDT 1997


* From the SCSI Reflector (scsi at symbios.com), posted by:
* PJohansson at aol.com
*
In a message dated 97-05-22 00:03:04 EDT, Mike_N_Bryan at notes.seagate.com
(Mike N Bryan) writes:

<<Peter J. this should probably go onto the SCSI reflector, also, but I'm not
on it.>>

You don't have to subscribe to SCSI at Symbios.com in order to be able to post
to it, Mike! And it's REALLY easy to subscribe (I think it's automated, a la
Majordomo). If you have any problems, drop a note to
John.Lohmeyer at Symbios.com. While you're at it, complain that the name of the
reflector isn't really representative of the standards group that uses it;
maybe we can persuade John to change it....

<<The dreq bit in the STATE_CLEAR and STATE_SET CSRs has the potential to
render 1394 devices inoperable. It allows any 1394 device to "disable
requests from unreliable nodes". Although IEEE 1212 states that devices may
"clear their own dreq bit when it has been impoperly set", that is not a
 reasonable solution. 1212 also states that support of dreq is not required
if bus-dependent mechanisms are provided to disable the requester's
initiation of bus transactions. If we assume that either of the two SBP-2
task management reset  functions, Target Reset and Logical Unit Reset, are
such mechanisms, then dreq support is not required of SBP-2 devices.
Ultimately, a 1394 Bus reset also meets this requirement.>>

One of the pleasures [sic] of working within the framework of inherited,
existing standards is just this sort of difficulty. The architects of 1212
didn't have SBP-2 devices in mind; I think the ability to clear one's DREQ
bit when improperly set is tantamount to refusing to let it be set in the
first place. Is there a problem?

<<Support by SBP-2 devices of other bits in the STATE_CLEAR and STATE_SET
CSRs should be discussed at the next Diskboys meeting in Irvine. My feeling
is that we should punt support of these CSRs.>>

It should definitely be on the agenda for the T10 ad hoc on Monday, June 16.
I'm not so sure we should "punt" so let's talk about solutions in Irvine.

Sincerely,

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