lkml.org 
[lkml]   [1996]   [Jul]   [2]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
SubjectRe: kerneld/request-route: An idea
Date
From
Alan Cox <alan@cymru.net> writes:
>You don't need a magic device you can do SLIP over a tty/pty pair to a
>monitoring daemon that the flips routes over, causes dials etc as
>appropriate. All in user space with no magic support needed. Being in user
>space diald also has some comprehensive filters added and can monitor the
>link to see when it looks idle and drop it.

This trend of reinventing the wheel makes me wonder just what is
wrong with the wheel we have. Obviously as the writer of the wheel
I'm a bit bothered by it, since I think the wheel is pretty good.
That said, I'm always willing to improve diald. So, let me ask,
exactly what is it about diald that bothers some people so much?

Is it the fact that it uses a SLIP link? The documentation? Something else?

Just as an aside. Now that the 2.0 kernel is here and we can properly
do route flipping in the stable kernel, I think I can rewrite diald
so that the SLIP link trick is unneeded. In particular I could just
monitor the dummy device with the SOCK_PACKET protocol. I'd need to
either use multiple dummy devices or IP aliasing in order to deal
with multiple possible source address (multiple copies of diald running),
but this shouldn't be a problem.
This approach would have the advantage that diald would then work for
all protocols, not just IP. Still no kernel changes needed though.

-- eric

---------------------------------------------------------------------------
Eric Schenk www: http://www.cs.toronto.edu/~schenk
Department of Computer Science email: schenk@cs.toronto.edu
University of Toronto




\
 
 \ /
  Last update: 2005-03-22 13:37    [W:0.117 / U:9.692 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site