> From: Jason Wang <jasowang@xxxxxxxxxx> > Sent: Tuesday, December 14, 2021 11:47 AM > > On Tue, Dec 14, 2021 at 12:11 PM Parav Pandit <parav@xxxxxxxxxx> wrote: > > > > > > > From: Jason Wang <jasowang@xxxxxxxxxx> > > > Sent: Tuesday, December 14, 2021 8:47 AM > > > > > > A question, consider the parent support both net and block, if user > > > use > > > > > > vdpa dev add name vdpa-a mgmtdev auxiliary/mlx5_core.sf.1 > > > > > > How do we know it's a net or block device? Or do you expect some > > > block specific attributes to be added? > > User should pass the virtio device id of type net/blk etc when creating the > device. > > No such in-kernel driver exists so far, so when it is added netlink interface > should be enhanced to accept the optional device id. > > So the plan is to reject the dev adding without specifying the device type? > Looks fine to me. Yes, this enhancement will be only for those mgmt. device that supports multiple device types. This ensures backward compatibility and also avoids the burden from user to not specify this for mgmt. device that support single type of device. _______________________________________________ Virtualization mailing list Virtualization@xxxxxxxxxxxxxxxxxxxxxxxxxx https://lists.linuxfoundation.org/mailman/listinfo/virtualization
