> I need to reacquaint myself with iWarp more, but I don't like > the idea of adding > CM calls as part of the verbs API, and in particular as part > of a generic RDMA > device structure. > > Does this suggest that each iWarp device driver will need to > implement a > connection establishment protocol? Isn't there a way to > generalize that into a > single iWarp CM module that can sit above multiple devices? How will > connections between different devices be supported? >
The Ammasso 1100 does do 100% connection setup. That's why we're pushing connection establishment verbs into the device struct. IMO, these functions are analagous to the process_mad function in the ib_device structs, which has no meaning to an iwarp device. So I think we have to admit up front, that the ib_device struct really has Infiniband-specific verb functions as well as iWARP-specific verb functions, and that's ok. (or maybe not :-) > I'm assuming that the Linux kernel will never permit an > established connection > to be offloaded onto a NIC. However it seems possible that a > new iWarp > connection could be done in a common way, with the result > passed into the device > through the modify QP call as the LLP stream. > > - Sean Assuming each RNIC supported some raw way to send and receive ethernet frames, then you could implement TCP, IP, ICMP, ARP etc al as a common stack to setup connections. I don't think we want to do this? BTW: We couldn't support this model with our 1100 card... Stevo. _______________________________________________ openib-general mailing list [email protected] http://openib.org/mailman/listinfo/openib-general To unsubscribe, please visit http://openib.org/mailman/listinfo/openib-general
