Obsolete mixed command and data, mixed data and response

Bob Snively Bob.Snively at Ebay.Sun.COM
Wed Nov 18 12:57:01 PST 1998


* From the T10 (formerly SCSI) Reflector (t10 at symbios.com), posted by:
* Bob Snively <Bob.Snively at Ebay.Sun.COM>
*

Gene,

The catalog is in my head.  The test of the catalog is the response on
the reflector, which asks the question to something like 500 people 
directly and 100s more indirectly.  So far, it is pretty close on this
subject.

I am not sure that there is any conflict with separating SPI and 
Fibre Channel in this particular case.

What the "Mixed command/data" and "mixed data/response" functions allow
is that in the same Information Unit (meaning "Sequence") the two elements
can be combined together.  The function says nothing about allowing two
separate sequences (an FCP_CMND sequence and an FCP_DATA sequence for example)
in the same loop tenancy, which is always possible and usually desirable, and
of course would not be effected by this change.

I believe that second implementation is more nearly akin to the SPI
implementation, where separate packet headers are provided for each
component in the transfer.

I believe it should have no effect on bridging, but this request to
obsolete the functions is intended to bring out any contrary opinions.

Bob




------------- Begin Included Message -------------


<<At present, there are no implementations that allow combining the command
and data in a single IU>>

     With each of these proposals it puzzles me where you find the catalog
of all implementations. Beyond that for this proposal I am wondering about
the impact of obsoleting what I think is a key feature of packetized SPI.
While it may be, if your catalog remains accurate, OK for fibre channel per
se, will it create a problem for bridging packetized SCSI through fibre
channel?

Gene

------------- End Included 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