> >  > I would like to propose a new project RDMA offload
 > >  > Framework and would like to request endorsement of the
 > >  > project by the networking, device drivers and HPC
 > >  > communities.
 > > 
 > > Could you expand a bit on the impact this project will have on the
 > > "core" stack (GLD/IP/TCP layers), if any?  I'm specifically concerned
 > > about increased complexity in those areas.
 > 
 > Meem,
 > 
 >       The project goal is to minimize number of changes to
 >       the core networking stack and number of interactions
 >       with the core networking stack. If we need any information
 >       from the core stack, we will try to get that information
 >       by querying the networking stack (for example: querying
 >       netinfo module).

Understood (and I'm happy to hear that), but could you expand on the
changes and interactions you anticipate that you'll need?

Thanks,
-- 
meem
_______________________________________________
networking-discuss mailing list
[email protected]

Reply via email to