[ofa-general] IPoIB CQ overrun

[Chris Elmquist]

On Tuesday (12/04/2007 at 05:15PM +0200), Eli Cohen wrote:
> 
> On Mon, 2007-12-03 at 17:40 -0800, akepner at sgi.com wrote:
> 
> > Looking at the contents of the CQ context table (right after the 
> > overrun at 15:50:53), do the producer and consumer indices look 
> > reasonable? I expected to find that producer_index + 1 == consumer_index.
> 
> I am not sure the value you get in the query for the consumer index
> reflects exactly what the consumer actually polled. I am checking with
> our FW team.
> Where you able to reproduce this on a system with less nodes?

It seems to be very difficult to reproduce when the node count is
small.  I am not sure we have yet.  On the other hand, very large
node counts make it real obvious.

Chris
-- 
Chris Elmquist          mailto:chrise at sgi.com      (651)683-3093
                        Silicon Graphics, Inc.     Eagan, MN