[openib-general] ipoib/mthca broken on ia64?

[Grant Grundler]

On Wed, Mar 23, 2005 at 12:16:08PM -0800, Roland Dreier wrote:
> It looks like the driver is working but the SM isn't bringing the
> ports to the active state.  The problem could still be on the host or
> the switch unfortunately.  What do you see in the files
> 
>     /sys/class/infiniband/mthca0/ports/2/counters/port_rcv_packets
>     /sys/class/infiniband/mthca0/ports/2/counters/port_xmit_packets

ionize:~# cat /sys/class/infiniband/mthca0/ports/*/counters/port_rcv_packets
0
0
ionize:~# cat /sys/class/infiniband/mthca0/ports/*/counters/port_xmit_packets
0
0

Ditto for iowa.

>     > Both systems are connected to a TS-90 switch (12 port?) which
>     > implements it own SM. Does this need a firmware upgrade maybe?
> 
> It's entirely possible.  Have you used the same switch successfully in
> the past?

Yes. This is the only switch I've got.

> You can check that the switch's SM is still working by
> connecting a cable directly from one switch port to another.  You
> should see both LEDs come on for both ports (possibly after a couple
> second delay) if the SM is still running.

I'm still at home and will test that after lunch when I head down
to the computer room (Cupertino).

> I believe the switch ports will get to INIT even with the embedded CPU
> completely hosed, so it's possible your switch has crashed or failed completely.

*nod*. I try the above first...then cycle power and see if it
comes back to life. The switch has been on since December or so.

thanks,
grant