<html><body>
<p>I have been following this and the other thread on CMA. There appears to be some opinions to removing the ib_at module and introduce CMA.<br>
Is that correct? <br>
<br>
True, CMA will need some form of address translation. Can we not use some incarnation of ib_at for that? I realize that ib_at has a net_device<br>
refcnt problem. Is this refcnt problem a usage issue rather than just a bug in the implementation? How would CMA solve the refcnt issue?<br>
What am I missing?<br>
<br>
Pradeep<br>
pradeep@us.ibm.com<br>
<br>
<br>
<tt>openib-general-bounces@openib.org wrote on 09/30/2005 07:27:32 AM:<br>
<br>
> On Tue, 2005-09-27 at 00:08, Michael S. Tsirkin wrote:<br>
> > Why does AT need to keep netdev reference for longer?<br>
> <br>
> I don't think it really does and could be changed. I think (but am not<br>
> sure) it was a convenience of implementation to try to make the netdev<br>
> reference counting simpler.<br>
> <br>
> It only needs to hold the netdev for sending the ARP (like SDP).<br>
> <br>
> It needs the underlying ib_device and port for ATS and path queries as<br>
> well as reregistration if the interface address changes and<br>
> deregistration if the IPoIB interface is removed. (SDP doesn't need to<br>
> worry about these aspects (only path queries).)<br>
> <br>
> -- Hal<br>
> <br>
> _______________________________________________<br>
> openib-general mailing list<br>
> openib-general@openib.org<br>
> <a href="http://openib.org/mailman/listinfo/openib-general">http://openib.org/mailman/listinfo/openib-general</a><br>
> <br>
> To unsubscribe, please visit <a href="http://openib.org/mailman/listinfo/openib-general">http://openib.org/mailman/listinfo/openib-general</a><br>
</tt></body></html>