Agenda:
Agenda:
1. Training
sequence/speed negotiation
http://www.t10.org/ftp/t10/document.05/05-397r3.pdf
The following ideas are for consideration.
Please investigate and discuss on the 6/15 call. Reflector traffic is welcome
and encouraged. Some concerns related to how the information is handled
and whether it reaches beyond the PHY layer.
Idea floated in the WG meeting of just sending
G1 traffic in the G3 window. This should get through without needing
training.
It would diverge from the normal window definition
by:
* Start with training pattern? (Suggest leaving
as idle time since speed will be determined later)
* Continue with ALIGN (0)/ALIGN (1) algorithm
like today
* End by sending information about supported
rates and features, not just sending ALIGN (1)s. Include:
- Additional rates supported (G3,
G4, G5, etc.)
- Whether the phy can receive SSC
- Advice to the other transmitter
to help it decide about tx amplitude, preemphasis, etc.
* "Receiver device"
characteristics (which it should know)
* System interconnect characteristics
(which it may know)
This would be the last trial window; there'd
be no need for subsequent (G4, G5, etc.) windows following the current
algorithm. The next window would be the final one, which would enable SSC
if supported, start with a training sequence, perform whatever is needed
for dword sync, etc. like has been discussed. If lock is unsuccessful,
the phys restart OOB. If lock is repeatedly unsuccessful, turn off
the fastest rate and try again. (Same issues as today)
2. Spread
Spectrum Clocking
a. SSC options / trade-offs [Witt]
http://www.t10.org/ftp/t10/document.06/06-267r0.pdf
Compares up-, center- and down-spreading
for SSC. Conclusion was that the short-term buffering impact of SSC is
minimal, but long-term impact was not covered. Makes a possible case for
center-spreading.
b. SAS-2 Spread-spectrum clocking by upspreading
(06-263) [Elliott]
http://www.t10.org/ftp/t10/document.06/06-263r0.pdf
Present cases and issues with different SSC
schemes. Supports up-spreading, but it looks like the center-spreading
calculation may have not been done correctly regarding the buffer dword
total. Kevin Witt will contact Rob about this.
PARTICIPANT INFORMATION:
Toll Free Dial in Number: (866) 279-4742
International Access/Caller Paid Dial In
Number: (309) 229-0118
PARTICIPANT CODE: 3243413
Webex information:
https://seagate.webex.com/seagate
Topic: SAS-2 PHY WG
Date: Thursday, June 15, 2006
Time: 10:00 am, Central Daylight Time (GMT
-05:00, Chicago)
Meeting number: 826 515 680
Meeting password: 6gbpsSAS
Alvin Cox
Seagate Technology, LLC
Tel 405-350-7424
Cell 405-206-4809
E-Mail alvin.cox@seagate.com