On Aug 6, 2014, at 2:48 AM, Samuel Ghinet <sghi...@cloudbasesolutions.com> wrote:
> Hello Nithin, > > I'd like to highlight two points here: > 1. the "datapath" on linux doesn't have all files Ovs-prefixed. I don't think > there's a specific requirement for datapath-windows for that. Sure, we don't have a requirement that the files should be prefixed with "Ovs". We just followed a convention - consistently. If you think it improves readibility if we don't have Ovs, I don't have any objections to it. > 2. the "datapath" on linux uses a lot of functionality of the linux kernel > (such as, the netlink protocol). > My belief is that the datapath-windows will, in time, need more things added > to it, things that the windows kernel API does not provide for us (one would > be, windows netlink protocol), which would inevitably make the file structure > more complex. Sure, if something is a self-contained entity and outside of the "core" OVS logic, it can reside in its own sub-directory. But, there has to be such code, first, is my belief. Netlink might be a candidate. If there are more than 3-4 files, it is better to put them in a separate directory. Similarly for the crypto library we talked about in another thread, to complete any IPSec offloads in the future. If it is a self-cotained entity outside of the core OVS logic, sure, it can go in its own sub-directory. thanks, Nithin _______________________________________________ dev mailing list dev@openvswitch.org http://openvswitch.org/mailman/listinfo/dev
