Skip site navigation (1)Skip section navigation (2)
Date:      Mon, 27 Feb 1995 23:27:32 -0800 (PST)
From:      Morgan Davis <root@io.cts.com>
To:        julian@tfs.com (Julian Elischer)
Cc:        tom@haven.uniserve.com, hackers@FreeBSD.org
Subject:   Re: pppd inactivity timeout?
Message-ID:  <199502280727.XAA05665@io.cts.com>
In-Reply-To: <m0rj0r4-0003vxC@TFS.COM> from "Julian Elischer" at Feb 27, 95 00:30:58 am

next in thread | previous in thread | raw e-mail | index | archive | help
Julian Elischer writes:
> 
> >This will use a lot of cputime.. a better way would be to call 
> timeout every 10 seconds or minute  while the line is up, and 
> have it decrement a counter.. if it reaches 0, you hang up.
> if not you just schedule another timeout..
> whenever you receive or send a packet, you just set the counter
> back to initial condition (maximum count)..
> that way you're doing a single write, not a whole function call(x2)
> for each packet.

How do netstat and pppstat get their information?  Isn't there a
counter somewhere in the kernel that keeps track of how many bytes
were sent and received?  If so, the task is even easier.  Set an alarm
to call a function once every so often (user-defined) and see if the
sent/received counts match.  If so, poof -- blow the connection off.
If they're different, do nothing and reset the alarm.



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199502280727.XAA05665>