About CD-ROM READ(10) Command

dallas at zk3.dec.com dallas at zk3.dec.com
Wed Apr 22 10:46:23 PDT 1998


* From the T10 (formerly SCSI) Reflector (t10 at symbios.com), posted by:
* dallas at zk3.dec.com
*

I have been following this thread with some interest (not a lot just some).

As a host vendor, I vote for option b, even though I would never issue
a command that does nothing.  The following are the reasons why I choose 
option b:
	
	1. Currently there is no wording of what happens for the
	   prefetch and media errors.  The host would want to know
	   about a prefetch failure. 

	2  Not deterministic of how many blocks/sectors to prefetch.

	3. Write(10) contains the same wording/meaning for Transfer
	   Length of 0.

	4  Pre-fetch behavior is already defined for the PRE-FETCH
	   command.

The PRE-FETCH command allows the host to specify starting where and how many
it would like in the cache.  The use of the PRE-FETCH command is 
determined by access pattern behavior. 

The use of option a for a prefetch and even the PRE-FETCH command itself
for read performance gains, itself is not very interesting.  This is due
to the OS/inter-connect/device command overhead, just might as well issue
the read. What is really needed is a new SCSI command (READ and PRE-FETCH).
This command would act as a 2 part command where you could specify what
you want for read data and prefetch parameters.  The device could satisfy
the read portion and return status for the read portion of the command.  
Then go off and do the prefetch (no status, other then device just tries
to do it).  This would save all the overhead associated with a second
command just to do a prefetch.

The above is not real interesting for disk but could have some value 
for DVDs. 

Thanks
Bill

------- Replied-To Message

Date: Tue, 21 Apr 1998 17:42:21 EDT
From: Gerry_Houlder%notes.seagate.com at us2rmc.enet.dec.com (MAIL-11 Daemon)
  To: t10 at symbios.com
Subj: Re: About CD-ROM READ(10) Command

* From the T10 (formerly SCSI) Reflector (t10 at symbios.com), posted by:
* Gerry_Houlder at notes.seagate.com
*
This brings up a question of what should be specified in the standard.
Clearly the T10 standard permits either action by the target. However, this
means that even a careful target designer could end up with a product that
doesn't conform to the expectations of the initiator. It also means that
targets are implemented with different firmware for different customers to
conform to their different expectations. Also note that option (a) is
defined as the PREFETCH command in the SCSI standard. If PREFETCH is
implemented, is it also proper to do the same thing for READ with zero
transfer length? Should PREFETCH be obsoleted if option (a) is the defined
resopnse for READ with zero transfer length?

The READ commands are defined separately in each command set document. Many
cases just refer to the SBC document for the command description. Perhaps
it is time to define a READ for CD ROM that defines the action in this
situation if it is not possible to get enough support to change the SBC
description.

If no further guidance is possible from the standard, It is probably safer
for a target to implement option (a). This will satisfy initiators that
expect this operation. It probably will not cause a problem for initiators
that expect option (b), because that option doesn't promise any particular
operation. What do the host vendors think of this?

>   Could you advice us that which interpretation is right about
> this command, (a) or (b) ?
>
>
>       (a) Even if transfer length=0, seek to Logical Block Address
>            and buffering to drive RAM from medium, then finish.
>
>       (b) Because of transfer length=0, NO seek operation, and
>            NO buffering from medium, then finish.


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

% ====== Internet headers and postmarks ======
% Received: from mail13.digital.com by us2rmc.zko.dec.com (5.65/rmc-22feb94) id
   AA09186; Tue, 21 Apr 98 17:38:26 -0400
% Received: from mpdgw2.symbios.com (mpdgw2.symbios.com [204.131.200.2]) by mai
  l13.digital.com (8.8.8/8.8.8/WV1.0d) with SMTP id RAA19680 for <Dallas at wasted
  .enet.dec.com>; Tue, 21 Apr 1998 17:38:20 -0400 (EDT)
% Received: (from root at localhost) by mpdgw2.symbios.com (8.6.8.1/8.6.6) id PAA1
  4561; Tue, 21 Apr 1998 15:29:54 -0600
% Received: from aztec.co.symbios.com(153.72.199.214) by mpdgw2.symbios.com via
   smap (V1.3) id sma014476; Tue Apr 21 15:29:34 1998
% Received: (from majordom at localhost) by Symbios.COM (8.6.8.1/8.6.6) id PAA1264
  8 for t10-outgoing; Tue, 21 Apr 1998 15:28:48 -0600
% Received: from mpdgw2.symbios.com (bastion.symbios.com [204.131.201.2]) by Sy
  mbios.COM (8.6.8.1/8.6.6) with ESMTP id PAA12641 for <t10 at symbios.com>; Tue, 
  21 Apr 1998 15:28:46 -0600
% From: Gerry_Houlder at notes.seagate.com
% Received: (from root at localhost) by mpdgw2.symbios.com (8.6.8.1/8.6.6) id PAA1
  4047 for <t10 at symbios.com>; Tue, 21 Apr 1998 15:28:47 -0600
% Received: from ns1.seagate.com(204.160.183.10) by mpdgw2.symbios.com via smap
   (V1.3) id sma013864; Tue Apr 21 15:28:22 1998
% Received: (from smap) by ns1.seagate.com  id OAA22135 for <t10 at symbios.com>; 
  Tue, 21 Apr 1998 14:28:19 -0700
% Received: from unknown(134.204.114.75) by ns1.seagate.com via smap (V1.3) id 
  sma022060; Tue Apr 21 21:27:44 1998
% Received: from sv-gw1.stsv.seagate.com (sv-gw1.stsv.seagate.com [134.204.14.9
  5]) by auth1.seagate.com  with SMTP id OAA04921 for <t10 at symbios.com>; Tue, 2
  1 Apr 1998 14:27:43 -0700
% Received: by sv-gw1.stsv.seagate.com(Lotus SMTP MTA SMTP v4.6 (462.2 9-3-1997
  ))  id 882565ED.0075DAAD ; Tue, 21 Apr 1998 14:25:34 -0700
% X-Lotus-Fromdomain: SEAGATE at INTERNET
% To: t10 at Symbios.COM
% Message-Id: <862565ED.00714657.00 at sv-gw1.stsv.seagate.com>
% Date: Tue, 21 Apr 1998 16:24:27 -0500
% Subject: Re: About CD-ROM READ(10) Command
% Mime-Version: 1.0
% Content-Type: text/plain; charset=us-ascii
% Sender: owner-t10 at Symbios.COM
% Precedence: bulk

------- End of Replied-To Message
*
* 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