Re: [EFM-Copper] Initialization of the PHY-Aggregation
As a related but different question since I missed the call...
It seems like the current draft is missing a way to statically provision the bonding groups, and disable the auto-discovery. Every other protocol that we do (LACP, OAM, PAUSE) can be enabled/disabled via Clause 30. I can't find a way in the current draft to say that I don't want to run auto-discovery, I just want these N ports to be an aggregated group.
Was this discussed on the call? Anyone else think this is a big "oops"?
-------Original Message-------
From: mathias.riess@infineon.com
Sent: 08/28/03 11:25 AM
To: stds-802-3-efm-copper@ieee.org
Subject: [EFM-Copper] Initialization of the PHY-Aggregation
>
>
Hi all,
as already outlined during the copper conference call last Tue (Aug.
26) I still see some ambiguities for the entire initialization of the PHY
aggregation using G.994. I understood the entire initialization process of
the PAF functionality in that way, that
1.it should be discovered which PMIs of a remote device might be
aggregated
(using the Npar(3) code points remote_discovery_register and the commands
'Get', 'Set if clear' and 'Clear if same'). Respective code points are
defined within the EFM tree of G.994.
2. the current aggregation functionality will be programmed on both -O and
-R side in the PMI_Aggregate_Registers.
I assume that these 2 actions need to be done during the G.994 sessions.
From my point of view Item 1 is already covered, but I still miss
respective
code points for item 2, for writing/reading the PMI_aggregate_register in
the G.994 EFM tree.
One suggestion would be that a new code point 'PMI_aggregate' will be
added.
If this bit is set, the Npar(3) code point remote_discovery_register will
be
interpreted as PMI_aggregate_register and the PMI_aggregate_register of
the
-r device will be written as described in section 61.2.2.7.3 (page 339
line
49 - page 340 line 8).
If we can agree on this procedure, a respective description needs to be
added to section 61.3.12.
Any feedback is welcome, may be some of the assumptions I stated above are
not correct.
Thanks
Mathias
>