Re: Specifying names for tap interfaces

[mlelstv%serpens.de@localhost (Michael van Elst)]

darrenr%NetBSD.org@localhost (Darren Reed) writes:

>If you're building scripts, be they for operation or installation
>then you need to "discover" which is the first interface. Is it
>fxp0, is it wm0, or is it hme0, etc, ceases to be an issue.

There isn't really much of a difference between enumerating all
interfaces eth0..N and just iterating through the interfaces
with ioctl(). [*]

You still don't know which interface is to be used for what purpose
if you just enumerate all interfaces. IMHO all the attempts like udev
cause more problems than they solve. The most useful naming scheme
for interfaces used to be one that describes topology or even physical
location, because then the purpose is usually clear from some cabling
scheme. But that also fails with hotplug devices.

Not even disks have persistent names. Volumes however do. And they
usually get autoconfigured. That's like saying you can identify
an interface by its IP address, and yes, that can be autoconfigured
too :)


[*] There should be a sysctl like hw.disknames for interfaces, so
you can do that easily in scripts.

-- 
-- 
                                Michael van Elst
Internet: mlelstv%serpens.de@localhost
                                "A potential Snark may lurk in every tree."