[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

RE: [ethmac] rx_ethmac



Hi,

   Phy don;t strip preamble bit's the reason for the shrink of the
preamble is due to repeters.

as for RMII just as MII require a full SFD meaning 1010_1011 the RMII
require the same meaing 10_10_10_11.

have a nice day

   Illan


-----Original Message-----
From: Kevin Kay [mailto:kevin.kay@espipd.com]
Sent: Sunday, December 23, 2001 9:36 AM
To: ethmac@opencores.org
Subject: RE: [ethmac] rx_ethmac


I noticed this as well when I looked at the receive MAC core state
machine.
I am not overly familiar with the specific standard, and just figured
that
it was for compatibility with any PHY's that may have the option of
stripping off the preamble bits (thus the start of a packet would only
be
the SFD nibble).

I did not concern myself with this issue too much since I am not
receiving
anything in my application.  I am only using the Tx MAC for my
application
and have stripped out the receive side code.  However, I do concur with
the
original poster's explanation.

Regards,
Kevin

> -----Original Message-----
> From: owner-ethmac@opencores.org [mailto:owner-ethmac@opencores.org]On
> Behalf Of yaron4fun@yahoo.com
> Sent: Sunday, December 23, 2001 6:15 AM
> To: ethmac@opencores.org
> Subject: RE: [ethmac] rx_ethmac
>
>
> Hi,
>
> From what i know about the ethernet fsm, i think ilan is right about
> the sfd and preamble. but in rmii when we detect "11", we should check
> that the 3 previos data di-bit where "10", only than we can asume
> that this is the sfd byte
>
> Thanks,
>  Yaron
>
> ----- Original Message -----
> From: "Illan Glasner" <iglasner@z... >
> To: <ethmac@o... >
> Date: Mon, 19 Feb 2001 16:44:56 -0800
> Subject: RE: [ethmac] rx_ethmac
>
> >
> >
> > 17-Feb-01
> >
> >    Hi,
> >
> >       Sorry for the mix word I guess this what happen when you
> > replay
> > and run a netlist check in the same time.
> >
> > Any way back to Premable and SFD the Pre Bye is define as 1010 and
> > 1010
> > and SFD Byte as 1010 and 1101
> >
> > since the first nibble in the SFD is like the nibble of the pre
> > than
> > many time the machine simple look for 1010 and we reffare to it as
> > pre
> > and thn for 1101 and we reffer to it as SFD
> >
> > BUT the MAC need to be able to find at least one nibble of 1010 and
> > than
> > 1101 before deciding that the packet has started.
> >
> > the rest of the preamable byte are need to be checked but not need
> > to be
> > presented and so your test bench need to test for different size of
> > preamable and this have to do with the "shrinking" of the preamable
> > when
> > the packet go through repeters.
> >
> > and so checking only for 1101 can make some mess in your MAC.
> >
> > When it come to transmite than you must generate 7 byte of pre and
> > 1
> > byte of sfd (well for specail purpose in point to point you can
> > reduce
> > this to increase the effective bandwidth, but this is not standart
> > so if
> > you do it make sure it is only an option).
> >
> > in respect to RMII and MII there is no differance between the two
> > the
> > RMII only come to solve pin count problem in high port number chip
> > and
> > the main logic differance is in the behave of rxdv.
> >
> > have a nice day
> >
> >    Illan
> >
> >
> > -----Original Message-----
> > From: Jamil Khatib
> >
> [/cgi-bin/post.cgi?cmd=new&to=jamilkhatib75%20at%20yahoo%20dot%20c
> om&msg=/ml-archive/archives/ethmac/0102/msg00009.shtml]
> > Sent: Monday, February 19, 2001 3:34 AM
> > To: ethmac@o...
> > Cc: mgalela@v...
> > Subject: RE: [ethmac] rx_ethmac
> >
> >
> > I do not know exactly what you are talking about but
> > as far as I know that the MAC should detect the
> > preamble and SFD and it should generate them on the
> > transmit side.
> > The whole bytes should be generated and checked
> > These notes are available at least on the MII and RMII
> > interfaces
> >
> > Regards
> > Jamil Khatib
> >
> > --- Mahmud Galela <mgalela@v... > wrote:
> > >
> > >
> > > I wonder what you are thinking about preamble and
> > > SFD (not
> > > SDF...just kidding) pattern.
> > > From previous e-mail (long time ago) i concluded
> > > that the SFD nibble can
> > > come from phy without preceeded by 1010.
> > > It means SFD can be located at first nibble or
> > > second niblle in a byte of
> > > data.
> > >
> > >
> > >
> > > Thanks
> > >
> > >
> > > Mahmud
> > >
> > >
> > >
> > > On Tue, 13 Feb 2001, Illan Glasner wrote:
> > >
> > > > 13-Feb-01
> > > >
> > > >    Hi,
> > > >
> > > > 	I belive your SDF detect is somewhat wrong.
> > > >
> > > > as soon as you see rxdv you start to look for SDF
> > > >
> > > > and your SFD is when you see 1101
> > > >
> > > > BUT the std specify SDF as both 0101 AND 1101
> > > togther and not only the
> > > > second nibble.
> > > >
> > > > this mean unless I'm mistaken and do correct me if
> > > I missed soemthing
> > > > but that if the packet come with let say junk
> > > nibble and one of the junk
> > > > nibble have 1101 than you machine will thing it is
> > > a real packet and
> > > > will start reading it.
> > > >
> > > > more over you might be not align as the 1101 might
> > > come in the first
> > > > nibble and not the second one.
> > > >
> > > > have a nice day
> > > >
> > > >    Illan
> > >
> >
> --
> To unsubscribe from ethmac mailing list please visit
http://www.opencores.org/mailinglists.shtml

--
To unsubscribe from ethmac mailing list please visit
http://www.opencores.org/mailinglists.shtml
--
To unsubscribe from ethmac mailing list please visit http://www.opencores.org/mailinglists.shtml