Skip site navigation (1)Skip section navigation (2)
Date:      Mon, 19 Feb 2001 23:34:12 -0800 (PST)
From:      Archie Cobbs <archie@dellroad.org>
To:        Poul-Henning Kamp <phk@critter.freebsd.dk>
Cc:        Luigi Rizzo <rizzo@aciri.org>, Archie Cobbs <archie@dellroad.org>, joe@tao.org.uk, wollman@khavrinen.lcs.mit.edu, cvs-committers@FreeBSD.org, cvs-all@FreeBSD.org
Subject:   Re: cvs commit: src/sys/netinet ip_fw.c ip_fw.h src/sbin/ipfw ipfw.8 ipfw.c
Message-ID:  <200102200734.f1K7YD505469@curve.dellroad.org>
In-Reply-To: <89834.982652391@critter> "from Poul-Henning Kamp at Feb 20, 2001 07:59:51 am"

next in thread | previous in thread | raw e-mail | index | archive | help
Poul-Henning Kamp writes:
> >> It would be somewhat slower because of the extreme modularity.
> >> But it would be more flexible, and possibly more robust and
> >> maintainable. Unfortunately it would also be a good deal of work.
> >> It would be a good exercise in networking programming though.. :-)
> >
> >... and maybe with a chance of improving existing code in the process.
> >As someone said, you don't get things for free, but...
> 
> SMPng will cost os more network performance than a NGification would.

Hmm.. considering the SMPng work we have to do anyway, NGification
wouldn't be so bad acutally.. :-)

Thinking further about this, tt would make sense for example to
consolidate all of our SMP-internal-kernel-packet-routing-among-subsystems
knowledge into one place, i.e., netgraph. Julian has already worked
out many issues with SMP and queueing, deadlock avoidance, locking,
etc. that could be leveraged and used more generally than just for
netgraph.

-Archie

__________________________________________________________________________
Archie Cobbs     *     Packet Design     *     http://www.packetdesign.com

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




Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200102200734.f1K7YD505469>