RE: [RFC] [PATCH 5/22 v2] [for 2.6.36] rdma/cm: update port reservation to support AF_IB

["Sean Hefty" <sean.hefty-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>]

>Right :) I just don't understand why you think AF_IB/RDMA_PS_TCP has
>to be a supported combination.

I don't see any reason to prevent it.  It works unless a specific check is added
in the kernel to prevent it.  Do I think it absolutely must be supported?  No -
so I don't have a strong disagreement on this.

It just seems like the code ends up as either:

	if (sa_family == AF_IB) {
		if (ps != PS_IB)
			fail;
		use_ps = get_ps(sid, mask);
	}
or

	if (sa_family == AF_IB) {
		if (get_ps(sid, mask) != ps)
			fail;
		use_ps = ps;
	}

>Is there some use case you see where having RDMA_PS_TCP/UDP work with
>AF_IB is an advantage?

Consider UDP, an application can create an id and join multicast groups using
AF_INET, AF_INET6, and/or AF_IB - whatever addressing scheme is most convenient.

What I have right now for user space support requires application changes to
make use of AF_IB anyway, so only supporting PS_IB isn't a big deal.  The
benefit to existing code is limited to user space path record queries. 

- Sean

--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html