[ofa-general] Re: IPOB CM (NOSRQ) [PATCH V8] patch

[Michael S. Tsirkin]

> Quoting Pradeep Satyanarayana <pradeeps at linux.vnet.ibm.com>:
> Subject: Re: IPOB CM (NOSRQ) [PATCH V8] patch
> 
> Michael S. Tsirkin wrote:
> >> @@ -815,7 +1168,9 @@ static struct ib_qp *ipoib_cm_create_tx_
> >>  	attr.recv_cq = priv->cq;
> >>  	attr.srq = priv->cm.srq;
> >>  	attr.cap.max_send_wr = ipoib_sendq_size;
> >> +	attr.cap.max_recv_wr = 1;
> >>  	attr.cap.max_send_sge = 1;
> >> +	attr.cap.max_recv_sge = 1;
> >>  	attr.sq_sig_type = IB_SIGNAL_ALL_WR;
> >>  	attr.qp_type = IB_QPT_RC;
> >>  	attr.send_cq = cq;
> > 
> > You never post a receive WR on this QP, do you?
> > So
> > 1. What's magic about 1 as max recv wr? Why not 0?
> > 2. If the remote sends a packet on this QP, it'llget closed,
> >    won't it? Looks like a spec violation.
> > 
> > 
> Good catch. I can probably set max_recv_sge to 0 too -right?
> I can do that in a separate patch later on.
> However, I see nothing in table 46 of the IB spec that tells me
> that it is a violation of the spec. Which section are you
> referring to?

The IPoIB RFC.

-- 
MST