[openib-general] Question

[Hal Rosenstock]

Hi Eitan,

On Tue, 2005-03-01 at 01:26, Eitan Zahavi wrote:
> Hal wrote:
> > So this looks like a workaround for a bug. Not sure what any of the
> other symptoms 
> > are but I'm real curious now. Can someone comment more on this ?
> 
> The ERR 3610 is really just a warning. It is caused by the Anafa1 chip
> responding with a LinearFDBTop 0xC000. 

Are you sure that the only problem when Anafa1 gets into this state ?
Does it continue to forward LR packets ? What happens with all the other
LIDs now theoretically in play ?

I also presume there's no fix for this with Anafa 1. Is that correct ?

> OpenSM does know how to handle that case and fix it.

Right, the workaround resets LinearFDBTop.

> > At a minimum, the SMA is reporting an invalid value for
> PortInfo::LinearFDBTop. I 
> > wonder if it also is incapable of forwarding DR MADs as well. That
> would explain 
> > this.
> There are no issues with that switch ability to do DR mads.

-- Hal