SMII help with Cisco 3750G and 2970G

[Guest]

Archived from groups: comp.dcom.lans.ethernet (More info?)

Anyone ever hear of the SGMII Gigabit Ethernet interfaces on these? They are
every other pair.

So,

SGMII == 3,4,7,8,11,12,15,16,19,20,23,24

RGMII == The rest of the ports

What is happening is, a LOT of desktop/server NIC's have issues connecting
or staying connected to Cisco's SGMII interfaces. Cisco's solution? Blame
the damn NIC and tell me to only use the RGMII ports. Pretty exensive 12
port switch I must say! This is happing on Intel PRO1000 CT and MT, Via
Rhine II 100Mbit, and Realtek 100Mbit NICs so far! Cisco briefly mentions
this ONLY for one NIC and blames it on Intel in CSCea77032 . Why the hell
are they selling a switch where half its ports are using a BARELY
IMPLEMENTED new standard, without putting a big red freaking stamp on these
ports? Instead, there is only an obscure mention in the release notes
relating to one particular NIC. TAC tells me it's only Intel, when I have
demonstrated other NIC's doing the same thing. The second I plug any of
these into one of the RGMII ports, all is well. Cisco's "fix" of forcing
speed/duplex has no effect! I am so freaking mad right now, I spent all day
on this. Anyone else hear of this? What is the solution?

 

[Guest]

Archived from groups: comp.dcom.lans.ethernet (More info?)

On Fri, 30 Jul 2004 00:07:50 -0700, "ccs" <temp@no.com> wrote:

>Anyone ever hear of the SGMII Gigabit Ethernet interfaces on these? They are
>every other pair.
>
>So,
>
>SGMII == 3,4,7,8,11,12,15,16,19,20,23,24
>
>RGMII == The rest of the ports
>
>What is happening is, a LOT of desktop/server NIC's have issues connecting
>or staying connected to Cisco's SGMII interfaces. Cisco's solution? Blame
>the damn NIC and tell me to only use the RGMII ports. Pretty exensive 12
>port switch I must say! This is happing on Intel PRO1000 CT and MT, Via
>Rhine II 100Mbit, and Realtek 100Mbit NICs so far! Cisco briefly mentions
>this ONLY for one NIC and blames it on Intel in CSCea77032 . Why the hell
>are they selling a switch where half its ports are using a BARELY
>IMPLEMENTED new standard, without putting a big red freaking stamp on these
>ports? Instead, there is only an obscure mention in the release notes
>relating to one particular NIC. TAC tells me it's only Intel, when I have
>demonstrated other NIC's doing the same thing. The second I plug any of
>these into one of the RGMII ports, all is well. Cisco's "fix" of forcing
>speed/duplex has no effect! I am so freaking mad right now, I spent all day
>on this. Anyone else hear of this? What is the solution?

BTW, where can one get a copy of the SGMII standard?

TIA,
Allan.

 

[Guest]

Archived from groups: comp.dcom.lans.ethernet (More info?)

Allan Herriman <allan.herriman.hates.spam@ctam.com.au.invalid> wrote in message news:<setjg01hkf534g2u26u6ab4v8l4js52peh@4ax.com>...
> On Fri, 30 Jul 2004 00:07:50 -0700, "ccs" <temp@no.com> wrote:
>
> >Anyone ever hear of the SGMII Gigabit Ethernet interfaces on these? They are
> >every other pair.
> >
> >So,
> >
> >SGMII == 3,4,7,8,11,12,15,16,19,20,23,24
> >
> >RGMII == The rest of the ports
> >
> >What is happening is, a LOT of desktop/server NIC's have issues connecting
> >or staying connected to Cisco's SGMII interfaces. Cisco's solution? Blame
> >the damn NIC and tell me to only use the RGMII ports. Pretty exensive 12
> >port switch I must say! This is happing on Intel PRO1000 CT and MT, Via
> >Rhine II 100Mbit, and Realtek 100Mbit NICs so far! Cisco briefly mentions
> >this ONLY for one NIC and blames it on Intel in CSCea77032 . Why the hell
> >are they selling a switch where half its ports are using a BARELY
> >IMPLEMENTED new standard, without putting a big red freaking stamp on these
> >ports? Instead, there is only an obscure mention in the release notes
> >relating to one particular NIC. TAC tells me it's only Intel, when I have
> >demonstrated other NIC's doing the same thing. The second I plug any of
> >these into one of the RGMII ports, all is well. Cisco's "fix" of forcing
> >speed/duplex has no effect! I am so freaking mad right now, I spent all day
> >on this. Anyone else hear of this? What is the solution?

I wasn't aware that SGMII was used for anything except between PHY and
MAC chips (or PHY to PHY connections) on a PCB. Assuming that is
true, I'm having a hard time understanding why a NIC on the other side
of a phy would cause a SGMII interface to not work, unless the SGMII
interface was not robust (which it sounds like it is not, since
multiple NICs don't work) - pretty much the only thing the NIC should
have to do is output 7 octets of preamble and a SFD. I suppose it's
possible the some NICs don't, but I'd be surprised. You don't need me
to tell you that it seems more likely this is a problem on the Cisco
side. You should press them for exactly what is wrong with the other
NICs, or else get a refund.

> BTW, where can one get a copy of the SGMII standard?

I believe that it used to be available on the Cisco web site, but I
spent a fair amount of time searching and can not find it there
anymore. I'll email a copy to you, Allan.

Marc

--
Reply to newsgroup - the From address is fake.

 

[novak]

Archived from groups: comp.dcom.lans.ethernet (More info?)

I would appreciate a copy of the spec as well. If possible please send
to novak_s@yahoo.com

Thanks,

Steve


--
novak
brought to you by http://www.wifi-forum.com/

 

[Guest]

Archived from groups: comp.dcom.lans.ethernet (More info?)

Is it possible for you to send me a copy of that spec as well?

Thanks,

Harry


--
hstello
brought to you by http://www.wifi-forum.com/

 

[Guest]

Archived from groups: comp.dcom.lans.ethernet (More info?)

-------------------------------------
Marc Randolph wrote:

> Allan Herriman <allan.herriman.hates.spam@ctam.com.au.invalid>
> wrote in message
> news:<setjg01hkf534g2u26u6ab4v8l4js52peh@4ax.com>...
>> On Fri, 30 Jul 2004 00:07:50 -0700, "ccs"
>> <temp@no.com> wrote:
>>
>> >Anyone ever hear of the SGMII Gigabit Ethernet interfaces on
>>> these? They are
>> >every other pair.
>> >
>> >So,
>> >
>> >SGMII == 3,4,7,8,11,12,15,16,19,20,23,24
>> >
>> >RGMII == The rest of the ports
>> >
>> >What is happening is, a LOT of desktop/server NIC's have
>>> issues connecting
>> >or staying connected to Cisco's SGMII interfaces. Cisco's
>>> solution? Blame
>> >the damn NIC and tell me to only use the RGMII ports. Pretty
>>> exensive 12
>> >port switch I must say! This is happing on Intel PRO1000 CT
>>> and MT, Via
>> >Rhine II 100Mbit, and Realtek 100Mbit NICs so far! Cisco
>>> briefly mentions
>> >this ONLY for one NIC and blames it on Intel in CSCea77032 .
>>> Why the hell
>> >are they selling a switch where half its ports are using a
>>> BARELY
>> >IMPLEMENTED new standard, without putting a big red freaking
>>> stamp on these
>> >ports? Instead, there is only an obscure mention in the
>>> release notes
>> >relating to one particular NIC. TAC tells me it's only Intel,
>>> when I have
>> >demonstrated other NIC's doing the same thing. The second I
>>> plug any of
>> >these into one of the RGMII ports, all is well. Cisco's
>>> "fix" of forcing
>> >speed/duplex has no effect! I am so freaking mad right now, I
>>> spent all day
>> >on this. Anyone else hear of this? What is the solution?

> I wasn't aware that SGMII was used for anything except between PHY and
> MAC chips (or PHY to PHY connections) on a PCB. Assuming that is
> true, I'm having a hard time understanding why a NIC on the other side
> of a phy would cause a SGMII interface to not work, unless the SGMII
> interface was not robust (which it sounds like it is not, since
> multiple NICs don't work) - pretty much the only thing the NIC should
> have to do is output 7 octets of preamble and a SFD. I suppose it's
> possible the some NICs don't, but I'd be surprised. You don't need me
> to tell you that it seems more likely this is a problem on the Cisco
> side. You should press them for exactly what is wrong with the other
> NICs, or else get a refund.

>> BTW, where can one get a copy of the SGMII standard?

> I believe that it used to be available on the Cisco web site, but I
> spent a fair amount of time searching and can not find it there
> anymore. I'll email a copy to you, Allan.

> Marc

> --
> Reply to newsgroup - the From address is fake.

Marc,
I´m working on a fpga design which requires gigabit connection, so I´m
thinking about using that interface, but I haven´t been able to find the
specifications.
Could you, please, email me a copy?
This is my email: gnathita@gmail.com

Thanks!






##-----------------------------------------------##
Article posted with Cabling-Design.com Newsgroup Archive
http://www.cabling-design.com/forums
no-spam read and post WWW interface to your favorite newsgroup -
comp.dcom.lans.ethernet - 2901 messages and counting!
##-----------------------------------------------##

 

[Guest]

Archived from groups: comp.dcom.lans.ethernet (More info?)

Take a look at
http://www.intel.com/support/network/adapter/pro100/sb/cs-007655.htm

It seems that Intel has a fix.


"gnathita" <gnathita_at_gmail_dot_com@foo.com> wrote in message
newsVSpe.10081$Yo5.9992@fe06.news.easynews.com...
>
>
> -------------------------------------
> Marc Randolph wrote:
>
>> Allan Herriman <allan.herriman.hates.spam@ctam.com.au.invalid>
>> wrote in message
>> news:<setjg01hkf534g2u26u6ab4v8l4js52peh@4ax.com>...
>>> On Fri, 30 Jul 2004 00:07:50 -0700, "ccs"
>>> <temp@no.com> wrote:
>>>
>>> >Anyone ever hear of the SGMII Gigabit Ethernet interfaces on
>>>> these? They are
>>> >every other pair.
>>> >
>>> >So,
>>> >
>>> >SGMII == 3,4,7,8,11,12,15,16,19,20,23,24
>>> >
>>> >RGMII == The rest of the ports
>>> >
>>> >What is happening is, a LOT of desktop/server NIC's have
>>>> issues connecting
>>> >or staying connected to Cisco's SGMII interfaces. Cisco's
>>>> solution? Blame
>>> >the damn NIC and tell me to only use the RGMII ports. Pretty
>>>> exensive 12
>>> >port switch I must say! This is happing on Intel PRO1000 CT
>>>> and MT, Via
>>> >Rhine II 100Mbit, and Realtek 100Mbit NICs so far! Cisco
>>>> briefly mentions
>>> >this ONLY for one NIC and blames it on Intel in CSCea77032 .
>>>> Why the hell
>>> >are they selling a switch where half its ports are using a
>>>> BARELY
>>> >IMPLEMENTED new standard, without putting a big red freaking
>>>> stamp on these
>>> >ports? Instead, there is only an obscure mention in the
>>>> release notes
>>> >relating to one particular NIC. TAC tells me it's only Intel,
>>>> when I have
>>> >demonstrated other NIC's doing the same thing. The second I
>>>> plug any of
>>> >these into one of the RGMII ports, all is well. Cisco's
>>>> "fix" of forcing
>>> >speed/duplex has no effect! I am so freaking mad right now, I
>>>> spent all day
>>> >on this. Anyone else hear of this? What is the solution?
>
>> I wasn't aware that SGMII was used for anything except between PHY and
>> MAC chips (or PHY to PHY connections) on a PCB. Assuming that is
>> true, I'm having a hard time understanding why a NIC on the other side
>> of a phy would cause a SGMII interface to not work, unless the SGMII
>> interface was not robust (which it sounds like it is not, since
>> multiple NICs don't work) - pretty much the only thing the NIC should
>> have to do is output 7 octets of preamble and a SFD. I suppose it's
>> possible the some NICs don't, but I'd be surprised. You don't need me
>> to tell you that it seems more likely this is a problem on the Cisco
>> side. You should press them for exactly what is wrong with the other
>> NICs, or else get a refund.
>
>>> BTW, where can one get a copy of the SGMII standard?
>
>> I believe that it used to be available on the Cisco web site, but I
>> spent a fair amount of time searching and can not find it there
>> anymore. I'll email a copy to you, Allan.
>
>> Marc
>
>> --
>> Reply to newsgroup - the From address is fake.
>
> Marc,
> I´m working on a fpga design which requires gigabit connection, so I´m
> thinking about using that interface, but I haven´t been able to find the
> specifications.
> Could you, please, email me a copy?
> This is my email: gnathita@gmail.com
>
> Thanks!
>
>
>
>
>
>
> ##-----------------------------------------------##
> Article posted with Cabling-Design.com Newsgroup Archive
> http://www.cabling-design.com/forums
> no-spam read and post WWW interface to your favorite newsgroup -
> comp.dcom.lans.ethernet - 2901 messages and counting!
> ##-----------------------------------------------##