[3111] in linux-net channel archive
Re: Persistent SLIP statistics
daemon@ATHENA.MIT.EDU (Michael Callahan)
Tue Jun 4 21:39:29 1996
Date: Tue, 4 Jun 1996 10:21:53 +0100 (BST)
From: Michael Callahan <callahan@maths.ox.ac.uk>
To: Matthias Urlichs <smurf@smurf.noris.de>
cc: linux-net@vger.rutgers.edu
In-Reply-To: <96Jun4.050504+0100met_dst.2222-534+1@work.smurf.noris.de>
On Tue, 4 Jun 1996, Matthias Urlichs wrote:
> In linux.dev.net, article <9605291552.AA15282@rsphy6.anu.edu.au>,
> Paul Gortmaker <gpg109@rsphy6.anu.edu.au> writes:
> >
> > Valid point -- the mapping of sl0 -> slN to various serial ports can
> > also vary depending on the order of usage, and thus the stats are not
> > necessarily bound to any one serial port. So I guess clearing the stats
> > for each up/down is actually the Right Thing To Do in this case.
> >
> How about clearing the stats for each _up_ event, but leave them alone for
> the _down_ event?
>
> That way, you can still collect the data when the interface goes down, yet
> there counters get cleared when a connection starts.
That's what PPP does.
Michael
