RE: [802.3ae] Comment about D3.1 48.2.4.5.1 "Sequence Q"
> -----Original Message-----
> From: THALER,PAT (A-Roseville,ex1) [mailto:pat_thaler@agilent.com]
> Sent: Wednesday, June 27, 2001 7:12 AM
> To: Boaz Shahar; HSSG (E-mail)
> Subject: RE: [802.3ae] Comment about D3.1 48.2.4.5.1 "Sequence Q"
>
>
> A change to this text should be editorial because the state
> machines specify
> the behavior and have precedence over the text.
>
> However, the suggested change text is not accurate either. Sequence
> ordered_sets are not always sent continuously on the XGMII.
> When they are
> generated by the 10GBASE-X sublayer, they will be sent
> continuously but when
> they are received having passed through a 10GBASE-X transmit
> process they
> will be sent as they are received (which will be once every 16 to 32
> columns).
Agree.
>
> When the 10GBASE-X receive process detects a fault, it sends
> the LF sequence
> ordered_set continuously to the XGMII. When the 10GBASE-X
> transmit process
> receives sequence ordered_sets, it sends a sequence ordered_set in the
> collumn that follows the column following the ||A||
> ordered_set. When the
> receive process receives a sequence ordered_set, it sends
> that sequence
> ordered_set.
Great. So, Pat, will you please propose this last statement instead of the
definition in D3.1 48.2.4.5.1?
Thx.,
Boaz
>
> Pat
> -----Original Message-----
> From: Boaz Shahar [mailto:boazs@mysticom.com]
> Sent: Monday, June 25, 2001 7:33 AM
> To: HSSG (E-mail)
> Subject: [802.3ae] Comment about D3.1 48.2.4.5.1 "Sequence Q"
>
>
>
> Hi,
> I think that the following change should be doen in the Draft:
>
> Instead of "Sequence ordered set are always sent in the
> column that follows
> an ||A|| ordered-set...etc"
>
> Write:
> "Sequence ordered set are always sent continuously on the
> XGMII (When fault
> is detected by the 10GBASE-X Receive process) and always
> sent in the column
> that follows an ||A|| ordered-set on the tx_code-group[39:0]"
>
> It may be considered editorial, thus preventing recirculation
> if necessary.
>
> Boaz
>