[openib-general] Re: RFC: cma: need rdma_unbind

[Michael S. Tsirkin]

Quoting r. Tom Tucker <tom at opengridcomputing.com>:
> Subject: Re: RFC: cma: need rdma_unbind
> 
> On Tue, 2006-04-25 at 00:27 +0300, Michael S. Tsirkin wrote:
> > Quoting r. Sean Hefty <mshefty at ichips.intel.com>:
> > > Subject: Re: RFC: cma: need rdma_unbind
> > > 
> > > Michael S. Tsirkin wrote:
> > > >I just tested this and this does not seem to be how it works:
> > > >if the socket is in TIME_WAIT state, SO_REUSEADDR let's you bind another
> > > >one to this port.
> > > >If the socket is not in TIME_WAIT state, you still get EADDRINUSE.
> > > >
> > > >socketfaq seems to agree
> > > >http://www.unixguide.net/network/socketfaq/4.5.shtml
> > > >
> > > >And CMA does not know about socket state - that's managed by ULP.
> > > 
> > > Would releasing the port from within rdma_disconnect() do what you need?
> > 
> > No, I don't want to disconnect yet - I am in the process of graceful
> > shutdown, I want to continue sending data.
> > 
> You can't (shouldn't be able to...) send data in TIME_WAIT. 

I should have said messages, not data.

> Are you trying to implement the TCP/IP half close semantics to pass some
> compatibility test? 

Yes.

-- 
MST