[openib-general] Re: CMA: compliancy issue?

[Michael S. Tsirkin]

Quoting r. Sean Hefty <mshefty at ichips.intel.com>:
> I think that the real issue here is that the CMA determines whether to hand 
> the REP to the user based on if it has a QP.  This was done to support 
> userspace. SDP wants the CMA to manage its QP, but still wants to see the 
> REP.

I'm a bit confused. What we currently have is that  CMA passes ESTABLISHED
event to SDP on REP. If so I think we can leave it as it is.

> What I need to determine is if it's easier/better for SDP to manage 
> the QP itself, or to change the CMA to expose the REP when it manages the 
> QP states.

I actually think all we have to do is to change CMA behaviour on REP:
send RTU after, and not before, calling user handler.
Since other ULPs don't seem t care when RTU is sent, they will continue
working.

No?

-- 
MST