[openib-general] Re: rdma_cm.h: comment nits.

[Michael S. Tsirkin]

Quoting r. Sean Hefty <mshefty at ichips.intel.com>:
> The intent is to keep connection establishment simple.  Socket users are 
> used to calling connect on the active side, and listen/accept on the 
>  passive side. The RDMA CM interface is consistent with that.

Sure. But you'll notice ESTABLISHED is something we have invented to map better
to asynchronous API. Its purely by choice that the same event is reporting
active and passive connections.  All I am saying is two different events for
active/passive side would be better.  It could be RESPONSE/ESTABLISHED or
ACTIVE_ESTABLISHED/PASSIVE_ESTABLISHED, or whatever.

What SDP does: looking at response private data and taking reject/accept
decisions based on this - is I think both portable and needed by other ULPs.
Note we don't need to pass any private data in the reject.

-- 
MST