[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: Intel PRO/1000 MT Dual Port Server Adapter Issues
- To: misc_(_at_)_openbsd_(_dot_)_org
- Subject: Re: Intel PRO/1000 MT Dual Port Server Adapter Issues
- From: Nick Bender <nbender_(_at_)_gmail_(_dot_)_com>
- Date: Tue, 18 Oct 2005 15:34:20 -0400
On 10/18/05, Bill Chmura <Bill_(_dot_)_Chmura_(_at_)_explosivo_(_dot_)_com> wrote:
> On Mon, 17 Oct 2005 22:10:53 -0400
> Jason Dixon <jason_(_at_)_dixongroup_(_dot_)_net> spake:
> > On Oct 17, 2005, at 8:59 PM, Damien Gardner Jnr wrote:
> > [ Redirecting back to misc@ where this belongs ]
> > > From: "Ken Gunderson" <kgunders_(_at_)_teamcool_(_dot_)_net>
> > > To: <pf_(_at_)_benzedrine_(_dot_)_cx>
> > >
> > >> I'm having some issues w/an Intel PRO/1000 MT Dual Port Server
> > >> Adapter
> > >> in an OBSD-3.7 firewall. The card is in the pci-x riser on one of
> > >> these puppies;
> > >> Dmesg complains the "The EEPROM Checksum Is Not Valid". I've tested
> > >> the card in a Winblows machine and it works. Any insights as to what
> > >> is going on here greatly appreciated.
> > >
> > > We had exactly the same problem with dual and quad versions of
> > > these same
> > > cards - we ended up swapping to PCI cards and changing the
> > > motherboard to
> > > something with enough PCI slots to support all the cards.. :\
> > If this is the case, I'm glad this became public. I'm about to fork
> > out for some of these. Can anyone suggest any specific Gig-E dual
> > port cards that work well with 3.7-3.8? Vendor recommendations welcome.
> > Thanks,
> > --
> > Jason Dixon
> > DixonGroup Consulting
> > http://www.dixongroup.net
> I have a Quad and Dual of these cards in the same machine - same as what
> Ken is showing and no problems. In a 1U system on a riser. Absolutely
> no problems with them.
> Not sure on the motherboard but I can send a full dmesg if you want one.
> Bill Chmura
> w. http://www.explosivo.com
What exactly does "The EEPROM Checksum Is Not Valid"? indicate?
I have three cheapo SuperMicro 1U servers (sm5013ct) which come
up with em0/em1 for the motherboard nics:
em0 at pci1 dev 10 function 0 "Intel PRO/1000MT (82541EI)"
rev 0x00: irq 10, address: 00:30:48:73:38:c6
em1 at pci1 dev 11 function 0 "Intel PRO/1000MT (82541EI)"
rev 0x00: irq 5, address: 00:30:48:73:38:c7
One of them will every three or four days drop the network and cough
up this message when you do a netstart. A reboot doesn't fix it but
a cold start usually will. The other two boxes have had no problems.
There are several messages in the archives relating to this including
one which suggests flashing the EEPROM using a SuperMicro utility,
which I have yet to try. I tracked it all the way to sys/dev/pci/if_em.c
but I'm afraid I've exhausted my capabilities at that point.
For now the bad box is the test box but I'm still curious...
Visit your host, monkey.org