SCSI ULTRA3: need help and advices

Gerry.Houlder at seagate.com Gerry.Houlder at seagate.com
Thu Aug 23 06:54:05 PDT 2001


* From the T10 Reflector (t10 at t10.org), posted by:
* Gerry.Houlder at seagate.com
*

The Write Buffer command with echo buffer option is used as part of a
domain validation algorithm. It is normal that most initiators will do a
few write buffer/ red buffer commands to verify that the cable/ backpanel
infrastructure can correctly pass data at the negotiated rate. I think the
initiator should react more gracefully to the illegal request response to
the Write Buffer command, but even a more graceful response would probably
limit the transfer rate to fast-40 or maybe even fast-20.

As a fellow developer of targets, my best advice is that it is far easier
for you to add support for the echo buffer option of Write Buffer command
than it will be to get the initiator to change its negotiation algorithm.
Remember that the standard is not a device procurement spec. Most buyers of
SCSI targets have procurement specifications that require support of many
features that are listed as "optional" in the standard. Also remember that
customer requirements do change over time. An item that was optional last
year may become mandatory this year even though the standard was not
changed (it typically is updated on two to three year cycles) and the
standard process is very slow about changing optional things to mandatory
things.





Rasmont Olivier <olivier.rasmont at barco.com>@t10.org on 08/23/2001 07:11:57
AM

Sent by:  owner-t10 at t10.org


To:   "'T10'" <t10 at t10.org>
cc:

Subject:  SCSI ULTRA3: need help and advices


* From the T10 Reflector (t10 at t10.org), posted by:
* Rasmont Olivier <olivier.rasmont at barco.com>
*
Hi,

I am developing a SCSI ULTRA3 interface for one of our products.
I try to establish a connection between an initiator an my device (=target)
in wide fast-80 DT mode. Everything is going correctly until the PPR
negotiation.
After that, the initiator requests a WRITE_BUFFER command in ECHO mode. My
device
does not support that command for the moment and so I reply with a
CHECK_CONDITION
and ILLEGAL_REQUEST sense key. Then the initiator restarts a PPR
negotiation
at lower data rate
followed by a WRITE_BUFFER again. And so on until ASYN mode.
What do I have to do to succeed to establish a FAST-80 protocol?
The WRITE_BUFFER command is not mandatory, is that normal that the
initiator
keeps going
using it, even after I reply with a CHECK_CONDITION?
Thanks for your help.



Olivier  Rasmont
Design Engineer
BARCO SILEX
Rue du Bosquet 7
1348 Louvain-la-Neuve
Phone:    +32 10 45 49 04
Direct:   +32 10 48 64 71
Fax: +32 10 45 46 36
http://www.barco-silex.com

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



*
* 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