iSCSI:SCSI Mode Pages

Gerry.Houlder at seagate.com Gerry.Houlder at seagate.com
Wed Sep 19 08:00:00 PDT 2001


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

I don't know if iSCSI targets are supposed to be different from SCSI
targets, but SCSI targets act like this:

(a) Separate mode pages are in effect for each LUN. Each LUN can have
different mode page values. There are two mode pages (both of which are
protocol specific) that break this rule, but the EMDP bit follows this
rule.

(b) The SCSI model places data buffers for each LUN in that LUN. Each LUN
manages its own command and data buffers as needed for its own use. There
is no data or command buffering at the "target" or "port" level or in the
interconnect layers. These layers may do routing of information to get them
to the correct LUN destination.

(c)  RAID devices the create virtual LUNs that are managed by a master
controller create their own implementation rules as they see fit, but the
result must look like the LUNs (a LUN in this case may be a collection of
storage space striped across multiple drives) obey the models described in
(a) and (b) above. Thus a RAID device may manage several LUNs and implement
the buffering for each LUN in its own buffer space but is has to look to
the outside world like each LUN controls its own space.





"Sanjeev Bhagat \(TRIPACE/Zoetermeer\)" <iscsi_t10 at sanjeevbhagat.com> on
09/19/2001 08:56:28 AM

To:   "Robert Snively" <rsnively at Brocade.COM>, <Gerry.Houlder at seagate.com>,
      <t10 at t10.org>, "IPS Refelctor" <ips at ece.cmu.edu>
cc:

Subject:  Re: iSCSI:SCSI Mode Pages



 I agree with Bob.So basically you would like to say that option no. 2
referred by me was correct. The data is stored in external buffers only and
not trasnferred to the SCSI LU untill the buffer is full with ordered data.

But now my question comes whether the iSCSI target will have a common
externa buffer and it will allocate a part of  buffer for each SCSI LU
inside or is the every SCSI LU going to maintain its own buffers.

Is it also possible for 2 different SCSI LUs to negotiate EMDP bit 0 and 1

Sincerely,

Sanjeev Bhagat

Tripace Europe

Tel # +31 (79) 361 2444

Mob# +31 6 24685051

Please, do visit our web site for more info:

http://www.tripace.com/

SCSI Chip Set(s), SCSI Host Adapters, SCSI Software Drivers to various
Operating Systems

SCSI, USB 2.0 & ATAPI/ATA Interconnectivity Chip Solutions

C R E A T O R S O F D A T A M O V I N G W O R L D S

This e-mail is confidential and may contain legally privileged information.
You should not disclose its contents to any other person. If you are not
the
intended recipient, please notify the sender above mentioned immediately.


----- Original Message -----
From: "Robert Snively" <rsnively at Brocade.COM>
To: <Gerry.Houlder at seagate.com>; <t10 at t10.org>
Sent: Wednesday, September 19, 2001 2:00 AM
Subject: RE: iSCSI:SCSI Mode Pages


> * From the T10 Reflector (t10 at t10.org), posted by:
> * Robert Snively <rsnively at brocade.com>
> *
> Note that in either case questioned by Sanjeev, it is the
> TARGET,  NOT the INITIATOR, that decides in what order the
> tranfers will be performed.
>
> In the case of a READ, EMDP = 1 allows non-contiguous setting
> of the Buffer Offset parameter.
>
> In the case of a WRITE, EMDP = 1 allows non-contiguous setting
> of the R2T Buffer Offset parameter.
>
> The target will never do such things unless it has
> committed to the proper manipulation of the data by allowing
> the EMDP bit to be set to a value other than 0.
>
> Bob Snively                        e-mail:    rsnively at brocade.com
> Brocade Communications Systems     phone:  408 487 8135
> 1745 Technology Drive
> San Jose, CA 95110
>
>
>
> > -----Original Message-----
> > From: Gerry.Houlder at seagate.com [mailto:Gerry.Houlder at seagate.com]
> > Sent: Tuesday, September 18, 2001 3:25 PM
> > To: t10 at t10.org
> > Subject: Re: iSCSI:SCSI Mode Pages
> >
> >
> > * From the T10 Reflector (t10 at t10.org), posted by:
> > * Gerry.Houlder at seagate.com
> > *
> >
> > The bit is intended to allow targets to send Modify Data
> > pointer messages,
> > which is the only mechanism to allow transmitting data out of
> > order. Data
> > that is not in order is required to be preceded with this
> > message, which
> > contains an offset to the data pointer to indicate where in
> > the data stream
> > the chunk of data belongs. This feature is considered a pain
> > to implement
> > and ususally isn't supported on most devices anyway.
> >
> > If you still think you want to try this out, generally the
> > Logical Unit is
> > required to buffer the data and rearrange it as needed (it
> > may not need to
> > rearrange the data, it will probably rearrange the transfer
> > to/from media
> > activity to get it organized right).
> >
> >
> >
> >
> >
> > "Sanjeev Bhagat \(TRIPACE/Zoetermeer\)" <iscsi_t10 at sanjeevbhagat.com>
> > @t10.org on 09/18/2001 01:30:03 PM
> >
> > Sent by:  owner-t10 at t10.org
> >
> >
> > To:   "IPS Refelctor" <ips at ece.cmu.edu>, "T10 Reflector" <t10 at t10.org>
> > cc:
> >
> > Subject:  iSCSI:SCSI Mode Pages
> >
> >
> > * From the T10 Reflector (t10 at t10.org), posted by:
> > * "Sanjeev Bhagat \(TRIPACE/Zoetermeer\)"
> > <iscsi_t10 at sanjeevbhagat.com>
> > *
> > I have a query regarding section 3.1.2 EMDP mode parameter.
> > Current text
> > says
> >
> > 3.1.2 E - Enable Modify Data Pointers Bit (EMDP)
> >
> >
> > ......
> >
> > If EMDP is set to 1, Data PDU sequences may be transferred in
> > any order. If
> > EMDP is set to 0, Data PDU sequences MUST be transferred
> > using continuously
> > increasing offsets.
> >
> >
> > Does this line mean that if EMDP=1 the out of order received
> > data PDUs by
> > the iSCSI target be transferred directly to the actual SCSI
> > LU (say a HDD)
> >
> > or
> >
> > does it mean that the if EMDP=1 the out of order received data PDUs be
> > stored in some external buffer provided by SCSI LU and
> > re-assembled and
> > ordered there before transferring it into the SCSI LU.
> >
> > Regards,
> >
> > Sanjeev
> >
> >
> > *
> > * 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
> >
> *
> * 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