> From: Jay Kuri <[EMAIL PROTECTED]>
> Date: 1999-07-08 11:19:01 -0700
> To: David Greenman <[EMAIL PROTECTED]>
> Subject: Re: Problem with fxp driver and 82559 cards
> Cc: [EMAIL PROTECTED]
> In-reply-to: <[EMAIL PROTECTED]>
> Delivered-to: [EMAIL PROTECTED]
> X-Loop: FreeBSD.ORG
>
> > >program... but if I go into the installer and try to transfer a
> > >distribution , it fails, locking in the same way. I'm talking
with Intel
> > >to see if they have had similar problems. I read something in
the source
> > >about the reciever has locked after garbage in the
syncronization bits...
> > >could it be a similar problem with the new chip, perhaps exposed by
> > >certain types of equipment?
> >
> > What's the cable plugged into?
> >
>
> We have tried:
>
> ATI switch (centrecom 3124tr) 10bT
> 3com SuperStack II (10/100)
> SMC TigerSwitch 10bT
>
> Same results on all of them. We've also tried setting media type
> to 10baseT/UTP in the 'ifconfig options' in sysinstall... no
difference.
Some off-the-wall thoughts. I haven't followed this closely, but
noticed the last few messages. Apologies if this is in the wrong
ballpark.
Have you tried a hub or a switch configured to stick with half-duplex?
Also, some switches "go dark" when there's a change in port status,
as might happen if a driver reinitializes the link. The switches are
executing a spanning tree algorithm on the new port, looking for
loops. If the switches can be configured to not do this (only needed
for switch-switch connection), you might try that as well. (I'm not
sure if the install issue is occuring soon after a boot, or is
something else).
Regards,
Justin
--
Justin C. Walker, Curmudgeon-At-Large *
Institute for General Semantics |
Manager, CoreOS Networking | When crypto is outlawed,
Apple Computer, Inc. | Only outlaws will have crypto.
2 Infinite Loop |
Cupertino, CA 95014 |
*-------------------------------------*-------------------------------*
To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-hackers" in the body of the message
