Skip site navigation (1)Skip section navigation (2)
Date:      Tue, 15 Dec 1998 09:52:05 +0100 (MET)
From:      hm@hcs.de (Hellmuth Michaelis)
To:        tlambert@primenet.com (Terry Lambert)
Cc:        current@FreeBSD.ORG, isdn@FreeBSD.ORG
Subject:   Re: if_sppp is BROKEN!!!
Message-ID:  <m0zpqD8-0000f8C@hcswork.hcs.de>
In-Reply-To: <199812150826.BAA16811@usr06.primenet.com> from Terry Lambert at "Dec 15, 98 08:26:39 am"

next in thread | previous in thread | raw e-mail | index | archive | help
>From the keyboard of Terry Lambert:

(Cc: stripped)

> Ideally, you would want an operation to propagate the full stack,
> up or down, as a result of one operation.

Can you please explain why one operation wants to propagate the full
stack ?

Background: In the current i4b ISDN implementation exactly this happens,
i.e. a layer 1 IRQ triggers an action in layer 2 triggering an action
in layer 3 which triggers an action (going down now) in layer 2 triggering
an action in layer 1 again. The problem of this going up and down at
IRQ time is not only the time lasting, but i.e. layer 2 being in a state
transition which is not fully finished when the reply from layer 3
going down enters layer 2 again (being still in the middle of the previous
up-going state transition).

Until now i was under the impression that only a decoupling of the layers
would make the transitions in one layer an atomic operation to avoid the
above described scenario.

hellmuth
-- 
Hellmuth Michaelis                                    Tel   +49 40 559747-70
HCS Hanseatischer Computerservice GmbH                Fax   +49 40 559747-77
Oldesloer Strasse 97-99                               Mail  hm [at] hcs.de
22457 Hamburg                                         WWW   http://www.hcs.de

To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-current" in the body of the message



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?m0zpqD8-0000f8C>