Re: cvs-src summary for June 21 - 28

From: Brooks Davis <brooks_at_one-eyed-alien.net>
Date: Tue, 29 Jun 2004 22:18:02 -0700
On Wed, Jun 30, 2004 at 02:03:33PM +0930, Thyer, Matthew wrote:
> > There's a small error in this pat of the summary.  It is not the case
> > that stf interfaces are created instead of stf0 interfaces in all cases.
> > Here's a table showing the changes.
> > 
> > Command		DEV	Version
> > ifconfig stf	stf0	old
> > ifconfig stf0	stf0	old
> > ifconfig 6to4	<fail>	old
> > ifconfig stf	stf	new
> > ifconfig stf0	stf0	new
> > ifconfig 6to4	6to4	new
> 
> Why is there a need to break backward compatibility?
> 
> Why cannot "ifconfig stf" result in "stf0" being printed and created?

I could probably change that part back, but it would complicate the code
for little or no benefit.  If a real user this change actually impacts
steps forward, I'll add the extra code too add behavior back.

The only reason stf has cloning at all is that it's a really nice
example of cloning without all the added complication of multiple
interfaces.  It does a very nice job of that.  When you get right down
to it, cloing stf devices is stupid because you can only have exactly
one of them.

> Does it make sense to have a device created called "stf", and if so
> why wasn't this required in the old implementation ?

As much sense as any name.  As long as you can configure the interface
with the tools you use, any interface name you want it fine.  Interface
names exist for admin convenience and as a way for cloning devices to
decide if they can create the interface the admin has requested.  They
mean nothing to the kernel.

-- Brooks

-- 
Any statement of the form "X is the one, true Y" is FALSE.
PGP fingerprint 655D 519C 26A7 82E7 2529  9BF0 5D8E 8BE9 F238 1AD4

Received on Wed Jun 30 2004 - 03:18:24 UTC

This archive was generated by hypermail 2.4.0 : Wed May 19 2021 - 11:37:59 UTC