VPD pages and multiple-protocol target ports

Gerry Houlder gerry.houlder at seagate.com
Wed Jul 31 14:39:51 PDT 2013


Formatted message: <a href="http://www.t10.org/cgi-bin/ac.pl?t=r&f=r1307314_f.htm">HTML-formatted message</a>

Hi Fred,
I initially thought the same thing you did, but then i realized that SAS
can define wide ports with multiple phys. Each phy could have a different
interface but all end up with the same port address.
On Wed, Jul 31, 2013 at 12:04 PM, Knight, Frederick <
Frederick.Knight at netapp.com> wrote:
>  I would expect them to be different ports.  The iSCSI port and the
FCoEport are connected to DIFFERENT SCSI domains.
> Consider EXTENDED COPY, for that to work, the two ports MUST have
> different relative port IDs.****
>
> ** **
>
> If the device receives a page 83h INQUIRY via the FCoE port, I would
> expect it to return info about the FCoE port.  If it receives a page 83h
> INQUIRY via the iSCSI port, I would expect it to return info about the
> iSCSI port.  If I’m a host and I send an INQUIRY command to what I know
> is an FC port looking for port specific information (such as the FC
> WWPortName), and I get back a response that includes iSCSI specific stuff
> (such as the SCSI name string), then I’m going to call that device broken.
> ****
>
> ** **
>
> I’m not quite sure how “the target port on which the INQUIRY command was
> received” can be understood to return information about two different ports
> (and FCoE port and an iSCSI port); given that the command would have been
> received over only one of those SCSI domain into only one of those two
> ports.****
>
> ** **
>
> Of course, as George pointed out, page 88h is different; it is supposed to
> return multiple different structures with multiple different types of
> information (for all the different ports – either known to the device
> server, or known to the whole target – depending on which LUN received the
> command).****
>
> ** **
>
>		  Fred****
>
> ** **
>
> *From:* owner-t10 at t10.org [mailto:owner-t10 at t10.org] *On Behalf Of
*Penokie,
> George
> *Sent:* Monday, July 29, 2013 11:40 AM
> *To:* Paul Hughes; t10 at t10.org
> *Subject:* RE: VPD pages and multiple-protocol target ports****
>
> ** **
>
> Paul,****
>
> ** **
>
> VPD page 83h may, but is not required to, return multiple transport
> protocols for a single SCSI target port. ****
>
> VPD page 88h should return all the transport protocols on all the SCSI
> target ports the logical unit that received the Inquiry command knows
> about. ****
>
> ** **
>
> Bye for now,****
>
> George Penokie****
>
> ** **
>
> LSI Corporation****
>
> 3033 41 St NW****
>
> Rochester , MN 55901****
>
> ** **
>
> 507-328-9017****
>
> george.penokie at lsi.com****
>
> ** **
>
> *From:* owner-t10 at t10.org [mailto:owner-t10 at t10.org] *On Behalf Of *Paul
> Hughes
> *Sent:* Wednesday, July 24, 2013 4:44 PM
> *To:* t10 at t10.org
> *Subject:* VPD pages and multiple-protocol target ports****
>
> ** **
>
> If a SCSI target device contains a target port that supports multiple
> transport protocols (e.g. an Ethernet port that supports iSCSI and FCoE),
> should the Device Identification (83h) and SCSI Ports (88h) VPD pages
> contain target port designation descriptors for all supported transport
> protocols, or just the transport protocol that was used to retrieve the VPD
> page?
> ****
>
> ** **
>



More information about the T10 mailing list