On Fri, Mar 08, 2019 at 08:18:01PM +0000, Jason Gunthorpe wrote: > On Fri, Mar 08, 2019 at 10:13:48PM +0200, Leon Romanovsky wrote: > > On Fri, Mar 08, 2019 at 02:16:31PM +0000, Jason Gunthorpe wrote: > > > On Wed, Mar 06, 2019 at 12:08:47PM +0200, Leon Romanovsky wrote: > > > > From: Leon Romanovsky <leonro@xxxxxxxxxxxx> > > > > > > > > Generalize the naming scheme for RDMA devices, so users will always > > > > see names based on topology/GUID information. Such naming scheme has > > > > big advantage that the names are fully automatic, fully predictable > > > > and they stay fixed even if hardware is added or removed (i.e. no > > > > reenumeration takes place) and that broken hardware can be replaced > > > > seamlessly. > > > > > > > > The naming policy is possible to chose from NAME_KERNEL, NAME_PCI, > > > > NAME_GUID or NAME_FALLBACK, which is controlled by udev rule. > > > > > > > > * NAME_KERNEL - don't change names and rely on kernel assignment. This > > > > will keep RDMA names as before. Example: "mlx5_0". > > > > * NAME_PCI - read PCI location and topology as a source for stable names, > > > > which won't change in any software event (reset, PCI probe e.t.c.). > > > > Example: "mlxp0s12f4". > > > > > > I don't think we should have the vendor/driver name in the stable > > > names. Ethernet doesn't do this. > > > > > > At worst it should be the base technology: > > > > > > ibp0s12f4 > > > rocep0s12f4 > > > iwp0s12f4 > > > opap0s12f4 > > > > It was my initial thought, but mlx4 with his dual mode, where one port > > can be IB and another can be RoCE doesn't allow us to use this scheme. > > Any idea how to name mlx4? > > It shouldn't be doing that in the first place, use the lowest port > number to set the name. I'm not convinced that base technology is right thing to do. Do we have clear (and not verbs) way to distinguish type of HCA? What should we do with non-RDMA devices like EFA or other big vendor who will invent "new, single vendor technology"? How can we recognize OPA? The fact that RDMA HCAs are not commodity allows us do not adapt netdev scheme and don't lose information which vendor name provides. If we want to adapt netdev naming, we won't accept EFA to RDMA as not standard technology. Are you ok with that? Thanks > > Jason
Attachment:
signature.asc
Description: PGP signature
