tech-kern archive

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Old Index]

Re: Interface description support



>>> Couldn't you just enforce the description to not match
>>> /[a-z]+[0-9]+/?
>> That breaks down if you want the descriptions to be "wan0", "lan0",
>> "lan1", etc.
> But you could use "wan-0", "wan_0" or "WAN 0" or whatever.

wan_0 and wan-0 are not fixes; _ and - are permitted in device names
(at least as of the config(8) source I have at ready hand; I haven't
checked -current).

> Or is there an argument that descriptions ought to be able to look
> like names?

I've been thinking about this.  Personally, I do not like forbidding
aliases (`descriptions') that match potential device names, but I have
been unable to come up with a clear use case; the closest I've found is
eth0, eth1, for Linux compatability, and that's a pretty weak case.

But it also occurs to me that calling it a "description" is misleading
if it's really an alternative name.  To me, "description" would be
something like "Inter-office VPN" or "4th floor desktops".  What's been
getting discussed here I would call an "alias", though I'd want to find
a different word for it because ifconfig(8)'s command line already has
an "alias" keyword.

Really, though, this whole thing is a botch.  Berkeley created a new
namespace for network interface names when they should have been just
nodes in /dev, and what we're discussing here is a pseudo-ln for that
other namespace.  The _right_ fix, it seems to me, is to move network
interfaces into /dev (possibly in a subdirectory?), where they should
have been all along, and then use ln, ln -s, mv, whatever, there.

/~\ The ASCII				  Mouse
\ / Ribbon Campaign
 X  Against HTML		mouse%rodents-montreal.org@localhost
/ \ Email!	     7D C8 61 52 5D E7 2D 39  4E F1 31 3E E8 B3 27 4B


Home | Main Index | Thread Index | Old Index