Re: control of order of inet devices

From: Mehmet Erol Sanliturk <m.e.sanliturk_at_gmail.com>
Date: Tue, 16 Apr 2013 12:03:14 -0700
On Tue, Apr 16, 2013 at 11:36 AM, Brooks Davis <brooks_at_freebsd.org> wrote:

> On Tue, Apr 16, 2013 at 10:08:00AM -0700, Adrian Chadd wrote:
> > Since people keep asking about this; maybe it's time we added a hint
> > to the bus code that allows for the unit to be set based on the pci
> > bus / slot / etc.
>
> I don't see how that would address Willy's request.  Neither the unit
> number or the if_index of an interface effects its order in
> getifaddrs() output.
>
> With modern bus hierarchies, you probably don't want to use the unit
> anyway as it loses too much information.  Some along the lines of
> Fedora's Consistent Network Device Naming would likely be more useful.
> That would be fairly easy to implement.
>
> -- Brooks
>


Fedora is using eth0 ,  ... , eth9 ,  but FreeBSD is using em* , re* , ..
names which is much better than Fedora ( or similar distributions )
approach .

In eth0 ,  ... , usage , it is necessary to record MAC addresses of the NIC
units ( because any name is not displayed like FreeBSD )  BEFORE starting
an installation to isolate which one is referenced : Even this is not very
useful sometimes because I am seeing they ( different Linux distributions )
are reporting MAC addresses differently for some NIC units , and no one of
them is identical to others to number eth with existing NIC units ( for
example , eth0 is not used for the same NIC by all the distributions , i.e.
, there is no any standard numbering scheme ) .

FreeBSD is using names ( Realtek , Intel , etc with other information also
) which is very easy to understand which NIC unit is referenced .

I think , in FreeBSD , the problem may occur when all of the NIC units are
the same model .
For such cases , there is a necessity to display also slot / port
information ( in that case MAC address is not showing slot / port
information ) .

Thank you very much .

Mehmet Erol Sanliturk
Received on Tue Apr 16 2013 - 17:03:21 UTC

This archive was generated by hypermail 2.4.0 : Wed May 19 2021 - 11:40:36 UTC