[SRP] Feb 15 Meeting Notes
cris.simpson at intel.com
Fri Feb 15 11:09:14 PST 2002
* From the T10 Reflector (t10 at t10.org), posted by:
* "Simpson, Cris" <cris.simpson at intel.com>
John Carrier Adaptec
Rob Elliott Compaq
Bob Nixon Emulex
Giles Frazier IBM
George Penokie IBM
Cris Simpson Intel
Ed Gardner Ophidian
Gerald Mauer QLogic
01-328r3 available at
1. 02-004 status
No work done or planned.
2. Review 02-064r3
Ed: Changed wording for UCSOLNT, added text for bit sense in several
GOP: Text needs to say "shall be set to one/zero"
Ed corrected, uploaded a new r3 to overwrite.
3. IBM #57
RobE: To case A, add "SRP_CMD".
Cover all SRP_ reqs.
After requesting that an RDMA channel be disconnected, after
being notified that an RDMA channel has been disconnected,
or upon receiving an SRP_I_LOGOUT request (see 6.5), an
SRP target port shall:
Not alter previously established conditions, including
MODE SELECT parameters, reservations, ACA,
and CA as a result of the disconnection.
A disconnect shall have no effect on SCSI parameters not specifically
related to the transport layer (e.g., reservations, most SCSI mode
pages, and SCSI log pages).
CRS: Change 'transport' to 'interconnect'
Behavioral differences w/ FC and iSCSI will make bridging
between SRP and those protocols difficult.
Add a mode page to control behavior?
RobE: We need to write up the bridging issues, so that we can talk to iSCSI
and FC people.
George & Rob will develop overview and suggestions. Likely a CAP
May require a protocol mode page.
Discussion of open technical comments
1. HP09: Security Protocol . . . . . . . . . . . . . . Pending
Reject, outside the scope of SRP.
2. HP27: Identifier construction rules . . . . . . . . . . Discussion needed
Anx B: Shall be an GUID (EUI-64?), which should be a CA GUID, plus an
Add a SRP-general description to Clause 5.
3. IBTA: IOControllerProfile I/O Class field . . . . . . Open
4. OD 3: Cross-channel reporting . . . . . . . . . . . . Discussion needed
Previously discussed, defer to SRP-2.
5. OD 8: Buffer formats & codes . . . . . . . . . . . . . Discussion needed
6. Troika: Correct Type Code in SRP_LOGIN_REJ . . . . . . Closed
7. IBM011: "This standard"
Sometimes you're discussing the protocol, not the specification.
In SRP, the protocol will be SRP, but the spec will be SRP-2.
8. IBM042: "Exactly once"
Change to: "once, without duplication"
9. IBM048: "Send tasks"
'd) For any outstanding SCSI tasks indicate to the application client that
the task has terminated with a service delivery system failure.'
Ed: Doesn't limit to the channel.
[Go with GOP's orignal suggestion.]
10. IBM097: Lack of bi-dir examples
5.4.2 Does say 'both'.
Need to say "No requirement that they be in same format."
Add at least one picture. May need a new heading.
11. intel0057: "byte address"
Byte offsets and lengths are pretty basic to the model.
SAM2 says "offset in bytes" - review to see if that works better.
12: edit002: "No requirements"
Add a footnote: "Does not change or override."
Feb 22 9000-1100 PST
Numbers: 1-888-316-5901, 1-617-801-9781
Cris Simpson 503.712.4333
Intel EPG/ACD Architecture Hillsboro, OR
PGP Fingerprint:0DA0 418E A27B 0B76 5F02 3DD4 0546 6D13 F88A 1E60
* 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