On Tue, Feb 23, 2016 at 05:01:31PM +0100, Jiri Pirko wrote: > Tue, Feb 23, 2016 at 04:55:28PM CET, go...@cumulusnetworks.com wrote: > >On Tue, Feb 23, 2016 at 03:45:51PM +0100, Jiri Pirko wrote: > >> Tue, Feb 23, 2016 at 03:34:19PM CET, go...@cumulusnetworks.com wrote: > >> > >> <snip> > >> >> > >> >> >> myhost:~$ dl port show > >> >> >> devlink0/1: type eth netdev ens4 > >> >> > ^^^^^^^^^^^ > >> >> >> devlink0/2: type ib ibdev mlx4_0 > >> >> > ^^^^^^^^^^^^ > >> >> >I think my only other question about this implementation is whether or > >> >> >not one would really want to have the true netdev/ibdev names mapped > >> >> >here. > >> >> > > >> >> >Would be as reasonable to simply specify the type (and there may be > >> >> >more > >> >> >types within ethernet that could be useful in multi-chip > >> >> >configurations) > >> >> >and then let normal infrastructure that exists today figure out how to > >> >> >map the names for the netdevs to the devices? > >> >> > >> >> What normal infrastructure you have in mind? There is no info about > >> >> devlink port mapping to netdev/ibdev anywhere. Only here. I might be > >> >> missing something but I fail to see what's wrong with it. > >> > > >> >I was simply wondering out loud if we _really_ wanted to name netdevs > >> >this way. I was suggesting that output could be like this: > >> > > >> >myhost:~$ dl port show > >> >devlink0/1: type eth > >> >devlink0/2: type ib > >> > > >> >mnd that udev/systemd/biosdevname/etc would take care of naming the > >> >device whataever it wanted. This appears to be essentially the same > >> >concern Hannes has. > >> > >> Wait. The only thing which will be renamed by udev is "devlink0". The > >> suffixes "/1" and "/2" are direct indexes as used inside the driver. > >> > >> And you need some link to netdev in case netdev exists - therefore > >> "netdev ens4" attribute is there. There's no other way to get the > >> mapping of "devlink0/1" to "ens4" anywhere else. > > > >So I think I had invisioned a slightly different workflow than what you > >just described. > > > >- Load PCI driver > >- Setup devlink attributes for your hardware > >- Create netdevs in network driver based on those attributes > > > >You don't need a netdev to reference any of the devlink specific > >parameters do you? > > No, I have a devlink handle and a devlink port index. That is enough.
That's what I read as well and why I wondered why you said this: > >> And you need some link to netdev in case netdev exists - therefore > >> "netdev ens4" attribute is there. There's no other way to get the > >> mapping of "devlink0/1" to "ens4" anywhere else. It just doesn't seem like referencing the netdev when performing set operations was needed. I see no issue having it appear a dump action after the fact, but it seems like devlink[unit number]/[port number] would be enough to reference the proper hardware mapping.
