[openib-general] [RFC] [PATCH 2/7] ib_multicast 2.6.20: add ib_multicast module to track join requests from the same port

[Sean Hefty]

Eitan Zahavi wrote:
> I agree that layering on top is easier. But does it really solve the 
> bug? I think not. If you would REPLACE the API and not provide both options
> (above and below refcount enforcement ) it would make sense to me.

We disagree on the philosophy here.  I view ib_mad as multiplexing to QP0/1, 
dispatching responses, and implementing rmpp.  I've tried to keep as much class 
specific information out of the mad layer as possible, and I'm resistant to 
changing that.  We don't want the mad layer implementing a bunch of class 
specific data.  Its not designed around providing that capability, nor do I 
think that it should be.

Consider that there's nothing that prevents a user from using the ib_mad 
interface to send CM mads directly, or even allocating a UD QP, and sending MADs 
by simply posting sends.

- Sean