For the change to the receive state machine, see comment 1159 which removes
the detection of |P| (which in d2.1 will be |Q|) which is the behavior Boaz
mentions. The decision to send constant ordered sets on the XGMII when
detecting a local fault was made. Comment 267 was part of that, but the full
resolution isn't in that comment's response. Also see comment 592 and 588.
These resolutions also impact clause 49 because it will now need to be able
to drop pulse ordered sets, but we didn't draft the text for that change at
the meeting. I have to do that.
From: Tim Warland [mailto:twarland@xxxxxxxxxxxxxxxxxx]
Sent: Wednesday, January 24, 2001 8:28 AM
Subject: Re: RF/LF
Boaz Shahar wrote:
Unfortunately I missed the discussions about RF/LF (Had to be in the XAUI
sessions). From the least I sow at the end, I thought that only the Tx
machine in the PCS-PMA of the DTE-XGXS considering any of the faults (In
order to come to a state where you transmit POS after /A/). Is this correct?
In the Rx direction you do not have to alternate the XGMII anymore-rather
you send the POS code whenever it comes, because the RS resets its machine
only after 128 non fault indicated bytes, and the max distance between POS
is 32 (Max A distance).
From my notes for clause 46, 48 and 49, I recall significant discussion
about LF/RF. From clause 46, the RF/LF detection and clearing
criteria are discussed in the response to comment 267 (per Steve
From Clause 48 and 49 we discussed the alternating LF (or RF)
with the idle pattern. At the time I thought we agreed that the
RS would no longer perform this alternating sequence. So in the
case of a LF, the RS sends a constant LF to the PCS which then
forwards LF to the line. In order to adapt the line rate, the PCS
layer could insert or delete LF's (or more correctly /LF/).
Did anyone else hear that? I can't find a comment response which
Tim Warland P.Eng.
Hardware Design Engineer Broadband Products
High Performance Optical Component Solutions
Nortel Networks (613)765-6634