> Wouldn't it make more sense to fix the bonding device or core networking
> (wherever the problem is) so that it recomputes the multicast list when
> the bonding hardware type changes?  After this patch, do we end up with
> an IPoIB interface that's not a member of the all hosts multicast group?
> (That seems like it would lead to confusing breakage later)
> 
>  - R.
You are right. The full fix to the example I brought needs to be higher in the 
stack.
I'll see how can I fix it there also.

But bonding is just an example (and I don't have any other yet). The fix in 
IPoIB 
prevents starvation of pending multicast addresses if one constantly fails to 
join.
Now, IPoIB assumes that if a join for one mcast group fails then all other 
joins would fail
or otherwise why not let groups a try? I think that this is a wrong assumption.

_______________________________________________
general mailing list
[email protected]
http://lists.openfabrics.org/cgi-bin/mailman/listinfo/general

To unsubscribe, please visit http://openib.org/mailman/listinfo/openib-general

Reply via email to