Question On Logical Unit Number field of SPI-4 SPI L_Q information n unit

Spalding, AlanC Alan_Spalding at adaptec.com
Fri Jun 14 12:49:48 PDT 2002


* From the T10 Reflector (t10 at t10.org), posted by:
* "Spalding, AlanC" <Alan_Spalding at adaptec.com>
*
Hi George,
Thanks for the reply, much appreciated. However, the SAM-2 Logical Unit
description (section 4.8) appears to conflict with your statement: 

>>
On the other hand, for devices that only contain a single LUN of zero it
makes no difference what the value of HiSUP is because the value in for the
LUN is all zeros in all cases.
<<


The SAM-2 spec (Rev 23) states:
"A logical unit number is a field containing up to 64 bits that identifies
the logical unit within a SCSI target device. If a SCSI target device
contains 256 or fewer logical units none of which are dependent logical
units (see 4.12) or extended addressing logical units (see 4.13), then its
Logical Unit Numbers shall have the format shown in table 1, which is a
single level subset of the format described in 4.12."

and

"When the single level subset format is used, the HISUP bit shall be set to
one in the standard INQUIRY data (see SPC-2)returned by logical unit 0."

According to these SAM statements: If a device contains only one LUN (LUN 0)
then the device contains "256 or fewer logical units" and SHALL use the
single level subset format, and "the HiSUP bit SHALL be set to one in the
standard Inquiry data".

Am I interpreting this incorrectly?

Also, one last question, this single level LUN field is defined as Byte 1 of
the LUN structure in SAM-2, this would be Byte 5 of the SPI L_Q IU in SPI-4,
correct? (Logical Unit Number starts at Byte 4).

Thanks in advance
alan     


-----Original Message-----
From: George Penokie [mailto:gop at us.ibm.com]
Sent: Friday, June 14, 2002 12:06 PM
To: Spalding, AlanC
Cc: t10
Subject: Re: Question On Logical Unit Number field of SPI-4 SPI L_Q
information n unit



Alan,
You ask some interesting questions.

1-According to what is stated today in SAM-2 if the target has either
dependent logical units or uses the single level LUN it would be required
to set the HiSUP bit to one. Because neither SAM-2 nor SPC-2 states
anything about the format of the LUN if the HiSUP bit is zero then anything
goes when it is set to zero. That is not a good thing and there are at
least two letter ballot comments on SAM-2 relating to that issue. So things
may change but I will not know until the next meeting in July.

On the other hand, for devices that only contain a single LUN of zero it
makes no difference what the value of HiSUP is because the value in for the
LUN is all zeros in all cases.

Also, there is no requirement in SPI-4 on the structure of the LUN field in
an L_Q IU.

2-REPORT LUNS is not the only way but any other way requires you to issue
INQUIRY commands to every LUN to see if it valid or not.

3-Right now it is vendor unique except that LUN zero is all zeros. Also,
right now, If the HiSUP bit is zero then anything goes regardless of the
LUN. If the HiSUP bit is one then either the single level LUN structure is
used or the hierarchical structure for dependent logical units is used. If
you look close you will see there is no difference between those two the
way they are defined today. Which, of course makes not sense. This will
also be looked into at the July meeting.

4-Yes the target has to use the same LUN it received in the command it the
L_Q IU responses to that command. If it does not then the initiator has no
idea what logical unit is responding.

Bye for now,
George Penokie

Dept 2C6  114-2 N212
E-Mail:    gop at us.ibm.com
Internal:  553-5208
External: 507-253-5208   FAX: 507-253-2880




 

                      "Spalding, AlanC"

                      <Alan_Spalding at ad        To:       George
Penokie/Rochester/IBM at IBMUS                                            
                      aptec.com>               cc:       "Spalding, AlanC"
<Alan_Spalding at adaptec.com>                                 
                                               Subject:  Question On Logical
Unit Number field of SPI-4 SPI L_Q informatio       n     
                      06/12/02 09:45 PM         unit

 

 

 




Hi George,
I hope you don't mind me emailing you directly (rather than posting to T10
reflector) but you seem to be the person most likely to respond to
questions
I post on the SCSI reflector. Also I believe that you were the author of
the
Addressing model for SAM proposal.


I have a few questions on the contents of the  Logical Unit Number field of
SPI-4 SPI L_Q information unit:

1) SPI-4 Rev 9 references SAM-2. SAM-2 states that if a device support 256
or fewer LUNs that it SHALL use the Single Level Lun structure format. Goes
on to state that in this format the HiSUP bit of Inquiry data SHALL be set.
Therefore, single Lun SCSI Disk drives supporting packetized transfers must
always have the HiSUP bit set to 1 in Inquiry data, correct?

2) Is using the REPORT LUNS command the only way to determine the number of
luns supported by a target? I.e. Set the Select Report field to 02 (all
luns) and divide the contents of the LUN LIST LENGTH field by 8.

3) What is the format of the SPI L_Q LUN field for devices that return
HiSUP
= 0 in Inquiry data? Is this Vendor unique? I can't find any information.
Does this mean that if a device supported greater than 257 Luns that a
value
of  00 00 00 00 00 00 00 0x where x is non-zero would be legal?

4) Are targets required to return the LUN field received from the Initiator
in the SPI L_Q(target to init) for a response even for illegal or
unsupported Lun values? I would think so, but I don't see this happening
today.

Thanks for your time (and I hope you don't mind the questions).

Alan

Alan C. Spalding
Intra-site work phone # 864-8369
Work phone # 949 455-8369
Work fax # 949 455-8406





*
* 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