|Thread Links||Date Links|
|Thread Prev||Thread Next||Thread Index||Date Prev||Date Next||Date Index|
We will need to propagate this fix also to the equivalent construct in the PD section, as well as the dual-sig PD section.
Given that the current scheme works and the only objection is that "it is not good practice", I'm not sure this is worth the effort.
From: Chris Bullock (bullock) <bullock@xxxxxxxxx>
Sent: Tuesday, March 14, 2017 16:23
To: Yseboodt, Lennart; STDS-802-3-4PPOE@xxxxxxxxxxxxxxxxx
Subject: RE: [802.3_4PPOE] Comment #95 and #346 homework
Lennart is correct that the original proposed fix introduced a race condition. I have an updated proposal that introduces a new variable to perform the hand-shaking. There is still no need for a new state. Please review the attached proposal, and let me know your thoughts.
Unfortunately the proposed fix also introduces a race condition.
I'm not sure we really need to fix this, but if we do, we need to avoid that race condition.
Agreed - adding a "done" exit condition is better than adding a state just to wait until done
George A. Zimmerman, Ph.D.
CME Consulting, Inc.
Experts in PHYsical Layer Communications
Looks like a clever fix.
On Mon, Mar 13, 2017 at 9:30 PM, Chris Bullock (bullock) <bullock@xxxxxxxxx> wrote:
It turns out that we did not need to add a “wait state” to the DLL state-machine since we already had a UCT exit condition from MIRROR_UPDATE that we could adjust. Please review the changes shown below to see if you agree that these changes satisfy the issue from comment #95 and #346.