[ofw] __cleanup_cep() oddity

Fab Tillier ftillier at microsoft.com
Sun Dec 6 17:00:19 PST 2009


Fab Tillier wrote on Sun, 6 Dec 2009 at 15:48:02

> Hi Sean
> 
> I think it has something to do with ND, but may be a relic of a time
> past when there was reference counting problems in the QP/CEP
> management when using ND.

Nope, not ND related - that's been there since well before ND came about...

> 
> I'll try to figure out what it's there for...
> 
> -Fab
> 
> Sean Hefty wrote on Sat, 5 Dec 2009 at 22:31:41
> 
>>> Seems like it is superfluous sanity check.
>> 
>> I was hoping that was the case, but I wasn't sure.  I added a print
>> inside the if statement and haven't seen it yet, but I'm only testing
>> winverbs at the moment.
>> 
>>> But it is better to set ASSERT and to return p_cep->ref_cnt.
>> 
>> I'll look at changing this.  I'm trying to debug a hang cleaning up
>> CEPs, so
>> I've been checking locations where the CEP ref_cnt is adjusted.
>> 
>> - Sean
>> 
>> _______________________________________________
>> ofw mailing list
>> ofw at lists.openfabrics.org
>> http://lists.openfabrics.org/cgi-bin/mailman/listinfo/ofw
> 
> 
> _______________________________________________
> ofw mailing list
> ofw at lists.openfabrics.org
> http://lists.openfabrics.org/cgi-bin/mailman/listinfo/ofw





More information about the ofw mailing list