[openib-general][RFC]: CMA IB implementation

[Hal Rosenstock]

On Wed, 2005-09-28 at 07:37, Michael S. Tsirkin wrote:
> Quoting Sean Hefty <mshefty at ichips.intel.com>:
> > Subject: Re: [openib-general][RFC]: CMA IB implementation
> > 
> > Guy German wrote:
> > > I believe that ib_at is still a valuable module even if ATS reverse ARP 
> > > is broken, and I think we should discuss this.
> > 
> > Here's my thinking on this.  ATS is broken as you mentioned for
> > reverse lookups.  However, if we want to keep ATS, I think that ATS
> > registration/deregistration should be integrated with IPoIB.  To keep
> > it separate, we will need to patch net_device to provide an rdma_ptr
> > as suggested by Roland.
> 
> I *think* that having rdma_ptr might be useful in its own right,
> as a nicer way to get at the ipoib private data, which we need anyway,
> dont we?
> I am not, however, sure what rdma_ptr should point to?
> Some kind of structure including ca, port and pkey?

In the case of IPoIB, it could point to the IPoIB netdevice struct and
have an IPoIB exported function to return these parameters (based on
rdma_ptr passed in) ?

> On a side note, I wander whether ATS can be split into a separate module
> so that people that dont need it can avoid loading it.

It is a separate module so I don't understand what you are saying.
The only people needing this are running kDAPL, uDAPL, or iSER
currently.

-- Hal