1394 S800BaseT Meeting Minutes - 2003 04 22

Attendees:

Michael Teener Apple teener@apple.com
Les Baxter Avaya les@baxter-enterprises.com
Colin Whitby-Strevens Apple Computer colinws@apple.com
Kevin Brown Broadcom kbrown@broadcom.com
Walter Hurwitz Broadcom whurwitz@broadcom.com
Scott Powell Broadcom spowell@broadcom.com
Richard Thousand Broadcom rthousand@broadcom.com
Kory Sefidvash Broadcom kory@broadcom.com
Burke Henehan Texas Instruments bhenehan@ti.com

Administration:

  1. Voting rules review. Person must attend 2 of last 3 meetings including the current meeting. One person, one vote. If a person has contributed substantially to the work product for a meeting, the 2/3 policy can be waived.
  2. Call for Patents. (see IEEE web site for details on IEEE patent policy);
    1. Michael Teener reviewed the statement sent to the reflectors about Apple's patent statement See attachment.
    2. Kevin Brown of Broadcom stated that Broadcom is pursuing patents that may be relevant to std and that Broadcom's intention was to be reasonable and non-discriminatory in its licensing. Broadcom will draft letter of assurance to that effect.
  3. There is now a web site: grouper.ieee.org/groups/1394/S800BASE-T/

Work Status:

Items that need to be finished

  1. Autonegotiation. Come up with proposal for auto-negotiation using "next" page and preferences for behavior for hubs and endpoints and all cases (1394 has peers, Ethernet has hubs and endpoints; endpoint to endpoint,endpoint to hub,hub to hub,etc) Assigned to Kevin Brown: See below
  2. Connection Management. What is process to establish the connection (1394b connection management),Assigned to Colin WS: Nothing To Report (NTR)
  3. Clocking and FIFO sizing for the reconciliation layers (what clocks, sharing FIFOs, etc),Assigned to Colin WS: Nothing To Report (NTR)
  4. PHY delay fields,Assigned to Colin WS: Nothing To Report (NTR)
  5. 1394b PHY changes. Leave scrambling in or out/ leave serializer in or out? Use simpler coder/decoder (8bit + 1 bit code or full 10 bits codes), Assigned to Mike JT
  6. Ethernet or IP bridging,Pending assignment to Peter J.: Nothing To Report (NTR)
  7. 1394.1 bridging or not.,Pending assignment to Peter J.: Nothing To Report (NTR)
  8. Make future proof to allow for higher data rates.: Nothing To Report (NTR)
  9. GigE changes,
  10. S100 over same pairs and negotiate correctly as 100 base TX
  11. Path for S100baseT
  12. Require a low power mode that does not go back to auto-negotiation, not what Gig E does today. (use 10baseT method 42 times a second to "tone"?)
  13. Progress/error reporting to PHY registers

Item "e" work:

Kevin Brown and Richard Thousand presented on changes required to the 1394b PHY to enable better implementation of the proposed Reconciliation layer. It is proposed to remove the 1394 scrambler and encoder. Instead have logic that adds two bits to the byte of data to identify whether that symbol is a control character, a request character, or a data character. The coding proposed is:
00 data
01 request
10 control
11 not defined

This will allow the logic to distinguish whether or not to push all symbols onto the FIFO. All data symbols need to be put into FIFO, however since request and control codes are repeated, only the first symbol need be put into the FIFO.

Discussion followed around does the FIFO always empty or will there be bits left over. There will be some bits left over, since FIFO pulls out a fixed amount every time. Then waits for fixed time before starting to pull data out again while sending idles on the connection. This will ensure that there are enough idle states to keep scrambler & other things synced up.

The proposal does not include the current 1394 clk resync FIFO, however it could be combined into one FIFO in theory. Do not have a perfect data rate match. Count on 1394b PHY portion to manage clock difference.

Item "a" work:

Walter Hurwitz then presented on Auto-Negotiation (A-N).

Auto-Negotiation (A-N) will probably require new Silicon. But want to not touch Gigabit Ethernet PHY, push the changes into the 1394 portion. Leave 1394b link interface the same and GMII interface to Gigabit Ethernet PHY the same (with exception of auto negotiation). As a goal if touch auto negotiation will want to try to put in power management using same primitives.

A-N Purpose: Find common capabilities How does it work: Only between 2 devices, exchanges 16bit data words that form a "base" page, and "next" pages, using 10 baseT, best common technology is selected. A-N ends once technology is enabled. One characteristic of current A-N is it can take seconds to change from one mode to another.

Early 10/100 baseT PHYs only do the "base" page A-N, not the extra pages.

There will not be a problem selecting between 1394a & 1394b, 1394a should never be connected to a RJ45 connector.

Inside the 802.3 standard clause 28 covers basic A-N. It is thought that only "Annex 28C Next Page Message Code Definitions" needs to be modified.

After discussion of what needs to be in S800 and capabilities for auto-negotiation it is thought either used the MC 8 page, which exists today and has "spare" bits may be used or the MC 9 page be used, which may not be supported by current silicon, but would not be opposed as much in the 802 committee. Action items were assigned to investigate if current silicon supports MC 9, if it does, then will go MC 9 route, if not, need more discussion to arrive at selected means.

What about using OUI as a vendor specific means of negotiation? Committee thinks it is too complex and wants to avoid.

Item "l" was then discussed:

In current 1394b S100b silicon the PHY must look for S100b 1394b toning to connect. Can 10baseT fast link pulses be used instead of 1394b tones? This seems to be the right track for doing connection management.

Discussion about the Ethernet MDIO,management data Interface, allows to read and write to registers inside the PHY to configure PHYs. It is part of the GMII interface. Also can be used to access proprietary registers. Need to read and write from 1394 side into recon layer with these interfaces to these registers.

Action Items:

AI 1: Action Item: Outline a 1394 PMD chapter in the fashion of 1394b to provide framework for insertion of reconciliation layer and changes required to 1394b PHY layer in to 1394b std. Assigned: Michael Teener

AI 2: Action Item: Do c-mode description of both reconciliation layer & 1394b PHY changes. Assigned: Colin Whitby-Strevens

AI 3,Check if can do MC-9 via SW with existing PHYs today? Assigned: Richard Thousand

AI 4,Based on AI 3, give proposal to 802.3 committee with our vision. Assigned: none

AI 5,Create proposal of low power mode signaling modes Assigned: Kevin Brown

Next meetings:

Apple Cupertino June 10 Starting 11:00 a.m.
1394 TA meetings Oxford July 8 Coordinate with Host Oxford Semi

Regards, Burke Henehan


stds-1394: Apple statement on patents, next meeting of S800Base-T study group


From: Michael Johas Teener<teener@apple.com>
Date: Wed, 9 Apr 2003 17:33:26 -0500
To: stds-1394@ieee.org
Subject: stds-1394: Apple statement on patents, next meeting of S800Base-T study group

One of my action items from the S800baseT meeting was to get an official statement from Apple on 1394 patents (in particular, any actions we are taking regarding the current effort). In record time, here it is:

"Apple protects its inventions with patent applications and other legal protections. Apple will license any of its FireWire patents required to meet an IEEE Specification on reasonable and non-discriminatory terms. Apple participates in the 1394LA licensing association as one convenient method of licensing all essential 1394 patents of Apple and other participants."
Remember, the next meeting is at Broadcom in Irvine, CA on April 22 from 10:30 until 15:00 ... Please RSVP to Kevin Brown (kbrown@broadcom.com).

--
Michael D. Johas Teener, Plumbing Architect, Apple
One Infinite Loop, MS: 305-2GM, Cupertino, CA 95014, USA
voice:+1-408-974-8512 fax:+1-408-974-9547 mailto:teener@apple.com
PGP ID 0x3179D202 http://xns.org/=Michael%20Teener
-------------------------- www.apple.com -------------------------