Subject: Re: PWS onboard tlp not working?
To: Wolfgang Rupp <rupp@chello.at>
From: Graham Allan <allan@physics.umn.edu>
List: port-alpha
Date: 07/19/2007 09:54:10
On Thu, Jul 19, 2007 at 07:54:18AM +0200, Wolfgang Rupp wrote:
> Am Mittwoch, den 18.07.2007, 19:09 -0400 schrieb Andrew Thompson:
> > I have the same problem with my PWS 500au. I managed to make it work
> > okay over a crossover to another machine, but it'd lock my hub up
> > identically to yours.
> 
> This is a common problem. You can work around it by setting a fixed
> mode and speed on the SRM console. I don't have my Alphas here at
> the moment, but IIRC, it is "set ewa0_mode [whatever]".
> 
> To get a list of valid options, give the "set ew[x]0_mode" command
> a bogus option.

This is provoking some vague memory...

I recall there being two possible transceiver boards in the PWS. I call
it this because I think the tulip chip is on the main i/o board, but
then there is a separate board holding the physical interface. One
model has just a 10baseT connector, and the other type also has coax
and AUI. I think the one which has 10baseT, coax and AUI doesn't
support auto-negotiation on the 10baseT port. The one which has only
10baseT does do auto-negotiate correctly.

There was the same situation with the DEC DE500 PCI cards. Some early
models didn't do auto-negotiation. I guess they were probably DE500-AA,
though I mainly remember them as having some orange component on the
board. Later DE500-BA (with no orange bits) did auto-negotiation just
fine.

I don't think the tulip was such a bad ethernet NIC - we never had any
real problems with them...

G.
-- 
-------------------------------------------------------------------------
Graham Allan - I.T. Manager - allan@physics.umn.edu - (612) 624-5040
School of Physics and Astronomy - University of Minnesota
-------------------------------------------------------------------------