On Mon, 19 Sep 2005, Guy German wrote:
> This is a draft of a generic cm abstraction layer implementation for > infiniband. > I would like to get your comments on it. > > Disclaimer: > ---------- > It is just a skeleton implementation *very* basically tested (it compiles). > > There are things not implemented yet, e.g: > - ib_cma_get_device > - ib_cma_get_src_ip > - some cm events cases > - arp retries (Maybe should implemented in at.c) > - protection from destroy while in callback (route/path) > - APM > > The implementation took reference of the the openib kdapl implementation, > therefore I added Mellanox and NetApp copyrights. If there are other > copyrights need to be added I will immediately do so. > > I am attaching the files to this mail and I will send them inlined in 2 > separate mails (one for the header and one for the c implementation). I'm glad to see discussion returning to this topic. Not to nit pick, but why did you use the prefix "ib_cma_" instead of "rdma_" or "rdma_cm_"? _______________________________________________ openib-general mailing list [email protected] http://openib.org/mailman/listinfo/openib-general To unsubscribe, please visit http://openib.org/mailman/listinfo/openib-general
