[systemd-devel] persistent network device names

[Keller, Jacob E]

On Mon, 2015-07-27 at 21:12 +0300, Andrei Borzenkov wrote:
> В Mon, 27 Jul 2015 17:52:59 +0000
> "Keller, Jacob E" <jacob.e.keller at intel.com> пишет:
> 
> > Hello,
> > 
> > I am working with a network device that can create virtual function
> > devices. When I create a large (>8) vfs for this device I get some
> > weird device names,
> > 
> > If I create 64 vfs, I see something like:
> > 
> > ens8
> > ens8f[1-7]
> > ens8s[1-7]
> > enp8s[1-7]f[1-7]
> > 
> 
> All those names come directly from kernel. udev does not invent them 
> -
> this is exactly what lspci shows. If it does not match lspci, you
> should provide more information.
> 
> > ens8f1-64 or something?
> > 
> 
> Then kernel should enumerate them so; you really need to discuss it
> there.

Hi,

So what I see from lspci is

08:00.0 Ethernet controller
08:00.1 Ethernet controller
08:00.2 Ethernet controller
08:00.3 Ethernet controller
08:00.4 Ethernet controller
08:00.5 Ethernet controller
08:00.6 Ethernet controller
08:00.7 Ethernet controller
08:01.0 Ethernet controller
08:01.1 Ethernet controller
08:01.2 Ethernet controller
08:01.3 Ethernet controller
08:01.4 Ethernet controller
08:01.5 Ethernet controller
08:01.6 Ethernet controller
08:01.7 Ethernet controller
08:02.0 Ethernet controller
08:02.1 Ethernet controller
08:02.2 Ethernet controller
08:02.3 Ethernet controller
08:02.4 Ethernet controller
08:02.5 Ethernet controller
08:02.6 Ethernet controller
08:02.7 Ethernet controller
08:03.0 Ethernet controller
08:03.1 Ethernet controller
08:03.2 Ethernet controller
08:03.3 Ethernet controller
08:03.4 Ethernet controller
08:03.5 Ethernet controller
08:03.6 Ethernet controller
08:03.7 Ethernet controller
08:04.0 Ethernet controller
08:04.1 Ethernet controller
08:04.2 Ethernet controller
08:04.3 Ethernet controller
08:04.4 Ethernet controller
08:04.5 Ethernet controller
08:04.6 Ethernet controller
08:04.7 Ethernet controller
08:05.0 Ethernet controller
08:05.1 Ethernet controller
08:05.2 Ethernet controller
08:05.3 Ethernet controller
08:05.4 Ethernet controller
08:05.5 Ethernet controller
08:05.6 Ethernet controller
08:05.7 Ethernet controller
08:06.0 Ethernet controller
08:06.1 Ethernet controller
08:06.2 Ethernet controller
08:06.3 Ethernet controller
08:06.4 Ethernet controller
08:06.5 Ethernet controller
08:06.6 Ethernet controller
08:06.7 Ethernet controller
08:07.0 Ethernet controller

but for some reason fedora22 shrinks "08:00:0" into ens8 instead of
"enp8s0f0". That is the most annoying part of this. I can live with it
if it were all consistent ie:

enp8s0f0, enp8s0f1 and so forth.

I don't think the kernel exports enough information to get "ens8f1-64",
but the fact that I get weird orders like

ens8 and then enp8s0 and then finally enp8s0f1-7 is what bothers me,
and *that* is definitely not directly ported out of the kernel.

Regards,
Jake