> -----Original Message----- > From: Jason Gunthorpe <jgg@xxxxxxxx> > Sent: Wednesday, February 20, 2019 3:57 PM > To: Doug Ledford <dledford@xxxxxxxxxx> > Cc: Parav Pandit <parav@xxxxxxxxxxxx>; Leon Romanovsky > <leon@xxxxxxxxxx>; Leon Romanovsky <leonro@xxxxxxxxxxxx>; RDMA > mailing list <linux-rdma@xxxxxxxxxxxxxxx> > Subject: Re: [PATCH rdma-next 5/5] RDMA/core: Add command to set > ib_core device net namspace sharing mode > > On Wed, Feb 20, 2019 at 04:53:59PM -0500, Doug Ledford wrote: > > On Wed, 2019-02-20 at 14:28 -0700, Jason Gunthorpe wrote: > > > On Wed, Feb 20, 2019 at 03:28:11PM -0500, Doug Ledford wrote: > > > > > > > Obviously, fixing the netlink would be the more robust solution. > > > > > > I think the way Parav has it works reasonably OK. The admin can > > > toggle the switch *before* creating any net namespaces and it is > > > perfectly atomic and non-disruptive. > > > > > > Trying to make it be a strong fence after net namespaces .. the best > > > we could probably do is to run over the client list and call a new > > > callback that should destroy anything outside the namespace. > > > > Or, alternatively, we just fail after namespaces have been created. I > > don't think it's out of reason to say that once namespaces are > > created, we can't change modes. > > I'm okay with that, easy enough to do. Parav? > Yes. that is a lot better option.
