Thread Links Date Links
Thread Prev Thread Next Thread Index Date Prev Date Next Date Index

RE: Chapter 46: preamble length

At 01:02 PM 03/28/2001 -0800, you wrote:
>The two illustrated cases are not all of the possible preamble shrinkage
>cases, they are the two extremes. 

Therefore, exactly like this should be defined for XGMII to cover everything
in-between else any deviation from it is a CHANGE, isn't it? Though, I am still
not clear about the NEED for preamble shrinkage in 10GE.

>For 1000BASE-X, there are two obvious
>receive cases, that illustrated in Table 35-4 of 7 preamble + SFD (no need
>for the transmitting PCS to align the /S/) and the unillustrated case of 6
>preamble + SFD (when the PCS needs to delay the /S/ to complete a /I/
>ordered set). 

PCS does not lie in MAC. PCS lies in PHY. In 1000BASE-X 
the interface is not GMII, it is Ten Bit Interface 8B/10B 
encoded and is not equaivalent to XGMII architecturally. 
This interface is close to XAUI. Anyway, there was a NEED
and the preamble was shrinked in 1000BASE-X. But that
is 1000BASE-X not 10GE. 

>/S/ is not defined at the GMII, it only exists at the PCS.

But /S/ is defined in XGMII interface replacing a preamble byte, isn't it a 
CHANGE from MII/GMII interfaces (which are equaivalent to
XGMII when you see the layered architecture)?

>The gigabit requirement is that MAC receive with any length of preamble just
>as was the case for 100 Mb/s and 10 Mb/s. Neither 1000BASE-X or 1000BASE-T
>PHYs produce the SFD only minimum case.

So going along these lines will 802.3ae keep the preamble definition 
across XGMII interface as was in 10/100/1000 MII/GMII inetfaces? 
Otherwise it is a CHANGE, right?


>--Bob Grow
>-----Original Message-----
>From: Sanjeev Mahalawat [mailto:sanjeev@xxxxxxxxx]
>Sent: Wednesday, March 28, 2001 12:19 PM
>To: Grow, Bob; 'Devendra Tripathi'; Grow, Bob; 'Danielle Lemay';
>Subject: RE: Chapter 46: preamble length
>The way it is specified in older versions at receive 
>across MII/GMII is a start of frame with just SFD and also 
>a start of frame with 7 bytes of Preamble and 1 byte of SFD.
>I do not see a preamble being /S/ + 2 bytes of Preamble (0x55)
>+ 1 byte of SFD (0xD5) across MII/GMII. So,  why do you think 
>/S/ + 2 bytes of preamble + 1 byte of SFD across XGMII 
>is NOT a CHANGE but keeping /S/ + 6 bytes of preamble 
>+ 1 byte of SFD is a CHANAGE and required?
>At 10:59 AM 03/28/2001 -0800, Grow, Bob wrote:
>>As Brad pointed out in his message, it would be a change to require that
>>preamble length be preserved.  Though generated by the MAC it's original
>>purpose was specifically for PHY layer requirements, not the MAC.  While
>>many DTE designs benefit from the extra memory accesses allowed by IPG and
>>preamble, no previous generation of Ethernet has assumed that seven
>>bytes would be received at the MAC.  This was true even with the
>>continuously clocked media of previous Ethernet speeds, where preamble
>>shrinkage was studied in the design and budgeted in the topology rules.  In
>>some Ethernet port types, the PHY only changes IPG lengths, in others, it
>>changes both (even for full-duplex, continuously clocked, gigabit
>>The review and comments that have produced the D3.0 text were simply to
>>preserve a characteristic of Ethernet MACs.
>>--Bob Grow
>>-----Original Message-----
>>From: Devendra Tripathi [mailto:tripathi@xxxxxxxxxxxx]
>>Sent: Wednesday, March 28, 2001 10:36 AM
>>To: Sanjeev Mahalawat; Grow, Bob; 'Danielle Lemay';
>>Subject: RE: Chapter 46: preamble length
>>I think Sanjiv has very good point. It is regressive to go back and make
>>preamble open
>>for PHY layers. After we have entered in continuous transmission mode
>>or data),
>>I do not see any reasoning of allowing "option" in preamble field.
>>Devendra Tripathi
>>VidyaWeb, Inc
>>90 Great Oaks Blvd #206
>>San Jose, Ca 95119
>>Tel: (408)226-6800,
>>Direct: (408)363-2375
>>Fax: (408)226-6862
>>> -----Original Message-----
>>> From: owner-stds-802-3-hssg@xxxxxxxx
>>> [mailto:owner-stds-802-3-hssg@xxxxxxxx]On Behalf Of Sanjeev Mahalawat
>>> Sent: Tuesday, March 27, 2001 6:08 PM
>>> To: Grow, Bob; 'Danielle Lemay'; stds-802-3-hssg@xxxxxxxx
>>> Subject: RE: Chapter 46: preamble length
>>> Hi Bob,
>>> At 02:21 PM 03/27/2001 -0800, Grow, Bob wrote:
>>> >
>>> >On transmit, a conforming implementation will send seven
>>> preamble plus the
>>> >SFD.
>>> >
>>> >On receive, there is no current function that will change that
>>> length, but
>>> >the concensus of the committee was to keep the option open.  (In
>>> 802.3z we
>>> >did change preamble length for idle alignment.)  The D3.0 text
>>> should make
>>> >it clear that an implementation should be tolerant to changes in
>>> >length, though it can still rely on lane alignment (Start in
>>> lane 0, SFD in
>>> >lane 3).  Text was added to warn that the Start and SFD could
>>> appear in the
>>> >same column.
>>> What is the reasoning behind letting a layer lower than
>>> MAC to touch the preamble?
>>> Since preamble is coded as data it belongs to MAC
>>> and no lower layer should be allowed to change
>>> and/or remove the length of preamble.
>>> Thanks,
>>> Sanjeev
>>> >
>>> >--Bob Grow
>>> >
>>> >-----Original Message-----
>>> >From: Danielle Lemay [mailto:dlemay@xxxxxxxxxxxxxxxxx]
>>> >Sent: Monday, March 12, 2001 10:38 AM
>>> >To: stds-802-3-hssg@xxxxxxxx
>>> >Subject: Chapter 46: preamble length
>>> >
>>> >
>>> >
>>> >
>>> >
>>> >Is it possible for the preamble+SFD to be less than 8 bytes ?
>>> >
>>> >thanks,
>>> >Danielle
>>> >
>>> >
>>> >
>>> >*******************************************
>>> >Danielle Lemay
>>> >Design Engineer, Nishan Systems
>>> >dlemay@xxxxxxxxxxxxxxxxx
>>> >408-519-3985
>>> >