[openib-general] [PATCH 1/2] perftest: enhancement to rdma_bw to allow use of RDMA CM

[Steve Wise]

On Wed, 2006-07-05 at 21:02 +0300, Michael S. Tsirkin wrote:
> Quoting r. Pradipta Kumar Banerjee <bpradip at in.ibm.com>:
> > Subject: Re: [openib-general] [PATCH 1/2] perftest: enhancement to rdma_bw to allow use of RDMA CM
> > 
> > Michael S. Tsirkin wrote:
> > > Thanks for following up on this, on the surface looks good.
> > > I'm busy at the moment, but a quick question:
> > > 
> > > Quoting r. Pradipta Kumar Banerjee <bpradip at in.ibm.com>:
> > >> +static void pp_wait_for_done(struct pingpong_context *ctx)
> > >> +{
> > >> +	struct ibv_wc wc;
> > >> +	int ne;
> > >> +
> > >> +	do {
> > >> +		usleep(500);
> > >> +		ne = ibv_poll_cq(ctx->rcq, 1, &wc);
> > >> +	} while (ne == 0);
> > > 
> > > I wander whether we can get rid of the rcq and an extra send at the
> > > end by simply closing the QP, and ignoring the possible completion with error
> > > for the last round. I think this would be a nice simplification.
> > > What do you think?
> > > 
> > Indeed we can get rid of the extra send at the end (pp_wait_for_done and 
> > pp_send_done can be removed). As for the rcq need to think about it. Its being 
> > used to synchronize the client and the server so that the client sends the first 
> > message.
> 
> Why do we need an RC send for this? CM or socket message should be enough, no?
> 

There's no CM message for this.  We could use the socket/tcp connection,
but the current patch doesn't event setup the TCP connection if you use
the --cma flag.  So the best and easiest thing to do is use SEND.  Why
is this an issue?  What's wrong with using the QP send/recv to sync up?

Steve..