[802.3_100GCU] Possible solutions for a bug in the LPI function

["Hugh Barrass (hbarrass)" <hbarrass@xxxxxxxxx>]

All 100G copper-people,

 

Sorry for the wide distribution.

 

Those of you in the know might have spotted that there’s a bug in the Receive LPI State Diagram (Fig 82-17). I’ve written down some notes on the problem and some possible solutions (see attached). . I can’t claim that this is the complete set of solutions, nor can I claim that any of them have been thoroughly thought through. Neither am I advocating any one of the solutions. I have sent this with the intention of spurring discussion in advance of the comment deadline so that we can avoid multiple competing solutions when we get to the May meeting.

 

I will sponsor an conference call to discuss this issue on Wednesday (4/24) at 8.00am PDT. I apologize in advance for anyone who can’t make it due to the scheduling or the short notice. Please let me know if you would like the call-in details or if you have any comments on the issue.

 

Note that the ballot an comment deadline is 4/30, so I encourage you to submit comments with detailed solutions (on any issue) – rather than hoping that others will find and fix the problems.

 

With regards,

 

Hugh.

 

Attachment: RxLPI_SM_01.pdf
Description: RxLPI_SM_01.pdf